Zhao et al. BMC Biology (2021) 19:2 https://doi.org/10.1186/s12915-020-00931-z

RESEARCH ARTICLE Open Access An updated tribal classification of Lamiaceae based on plastome phylogenomics Fei Zhao1†, Ya-Ping Chen1†, Yasaman Salmaki2, Bryan T. Drew3, Trevor C. Wilson4, Anne-Cathrine Scheen5, Ferhat Celep6,7, Christian Bräuchler8, Mika Bendiksby9,10, Qiang Wang11, Dao-Zhang Min12, Hua Peng1, Richard G. Olmstead13,BoLi12* and Chun-Lei Xiang1*

Abstract Background: A robust molecular phylogeny is fundamental for developing a stable classification and providing a solid framework to understand patterns of diversification, historical biogeography, and character evolution. As the sixth largest angiosperm family, Lamiaceae, or the mint family, consitutes a major source of aromatic oil, wood, ornamentals, and culinary and medicinal herbs, making it an exceptionally important group ecologically, ethnobotanically, and floristically. The lack of a reliable phylogenetic framework for this family has thus far hindered broad-scale biogeographic studies and our comprehension of diversification. Although significant progress has been made towards clarifying Lamiaceae relationships during the past three decades, the resolution of a phylogenetic backbone at the tribal level has remained one of the greatest challenges due to limited availability of genetic data. Results: We performed phylogenetic analyses of Lamiaceae to infer relationships at the tribal level using 79 protein-coding plastid genes from 175 accessions representing 170 taxa, 79 genera, and all 12 subfamilies. Both maximum likelihood and Bayesian analyses yielded a more robust phylogenetic hypothesis relative to previous studies and supported the monophyly of all 12 subfamilies, and a classification for 22 tribes, three of which are newly recognized in this study. As a consequence, we propose an updated phylogenetically informed tribal classification for Lamiaceae that is supplemented with a detailed summary of taxonomic history, generic and species diversity, morphology, synapomorphies, and distribution for each subfamily and tribe. Conclusions: Increased taxon sampling conjoined with phylogenetic analyses based on plastome sequences has provided robust support at both deep and shallow nodes and offers new insights into the phylogenetic relationships among tribes and subfamilies of Lamiaceae. This robust phylogenetic backbone of Lamiaceae will serve as a framework for future studies on mint classification, biogeography, character evolution, and diversification. Keywords: Lamiaceae, Lamioideae, Mints, Phylogenomics, Tribal relationships

* Correspondence: [email protected]; [email protected] †Fei Zhao and Ya-Ping Chen contributed equally to this work. 12Research Centre of Ecological Sciences, College of Agronomy, Jiangxi Agricultural University, Nanchang 330045, China 1CAS Key Laboratory for Plant Diversity and Biogeography of East Asia, Kunming Institute of Botany, Chinese Academy of Sciences, Kunming 650201, China Full list of author information is available at the end of the article

© The Author(s). 2021 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Zhao et al. BMC Biology (2021) 19:2 Page 2 of 27

Background gynoecial structure. Although a deeply four-lobed ovary Lamiaceae, generally known as the mint family, have with a gynobasic style is typical for most traditionally long been known for their aromatic oils, which have recognized Lamiaceae (i.e. Lamiaceae s.s.), and an played an undeniably significant role within culinary, unlobed ovary with a terminal style is typical of most medicinal, and horticultural aspects of human history. Verbenaceae, there exists in both families a continuum Species of Lamiaceae are of wide economic importance in extent of lobing and separation of fruits into single as sources of wood (e.g., Tectona grandis L. f.), landscape seeded units [9]. Noting this, Cantino [9, 10] carried out ornamentals (e.g., scarlet sage [Salvia splendens Sellow a cladistic analysis of the Lamiaceae s.s. and the Verbe- ex Wied-Neuw.]), cosmetics (e.g., lavender [Lavandula naceae s.l. based on 85 morphological and anatomical angustifolia Mill.]), culinary herbs (e.g., basil [Ocimum characters, which provided support to reject that the basilicum L.], oregano [Origanum vulgare L.], thyme Lamiaceae s.s. was monophyletic, demonstrating several [Thymus vulgaris L.]), and medicinal herbs (e.g., Korean clades of the Verbenaceae s.l. recovered among clades of mint [Agastache rugosa (Fisch. & C.A. Mey.) Kuntze], the Lamiaceae s.s. Based on these results, Cantino et al. peppermint [Mentha × piperita L.]). Despite the recogni- [11] published a list of subfamilies and genera of the tion of this family (Lamiaceae s.s.) from advances in sys- Lamiaceae s.l. that had been proposed earlier by Junell tematics and taxonomy of the late twentieth century, the [12]. This incorporated the transfer of the cymose sub- family has historically been considered a “natural” group families Caryopteridoideae, Chloanthoideae, Viticoideae, based on a combination of readily recognizable features Symphorematoideae, and tribe Monochileae to the ex- such as an herbaceous habit, quadrangular stems, oppos- panded Lamiaceae, rendering the Verbenaceae s.s. as ite phyllotaxy, bilabiate flowers, a gynobasic style, and only the subfamily Verbenoideae. Verbenaceae s.s. can four nutlets. However, morphological and molecular be recognized by having racemose inflorescences, tricol- phylogenetic studies in the past three decades have sig- porate pollen, and ovules attached to the carpel margins, nificantly changed the concept of the family, and an ex- while the Lamiaceae s.l. generally possess thyrsoid inflo- panded Lamiaceae (Lamiaceae s.l.) is now widely rescences, colpate pollen, and ovules attached to the accepted. As currently circumscribed, Lamiaceae com- sides of the false septa of ovary [13]. Moreover, the Ver- prise more than 230 genera and over 7000 species, mak- benaceae s.s. have thickened stigma lobes with conspicu- ing it the sixth largest angiosperm family and the largest ous stigmatic tissue, hypocrateriform corollas with family in the order Lamiales [1–3]. Although unequivo- included stamens, and usually terete stems, whereas in cally shown to be members of the family, inclusion of the Lamiaceae s.l., stigma lobes are slender with incon- some disparate groups such as Vitex L. (originally placed spicuous stigmatic tissue, corollas that are rarely hypo- in Verbenaceae because they were trees with fleshy crateriform, and stems are typically quadrangular. Since fruits) has challenged the earlier concepts of the family. Cantino et al. [11], the expanded concept of the Lamiaceae Early infrafamilial classifications within Lamiaceae s.l. has been consistently supported as monophyletic by were predominately based on the treatment of Bentham molecular phylogenetic studies [14–20] and is widely ac- [4], who divided the family into eight tribes. Briquet [5], cepted in various classifications [1, 3]. We acknowledge for example, followed the division of Bentham [4], but these results and use the names Lamiaceae and Verbena- raised some of the tribes to subfamilial rank and merged ceae in their contemporary circumscription. Though today four tribes into the single large subfamily Lamioideae. both Lamiaceae and Verbenaceae are placed within “core Erdtman [6], however, recognized only two subfamilies Lamiales” of the asterids, they have unexpectedly not been based on palynological distinctions, viz., Lamioideae recovered as sister taxa despite their morphological simi- (with tricolpate pollen shed at the two-celled stage) and larities: Lamiaceae belong to a clade that includes Maza- Nepetoideae (with hexacolpate pollen shed at the three- ceae, Phrymaceae, Wightiaceae, Paulowniaceae, and celled stage). Combining the classifications of Briquet [5] Orobanchaceae, whereas Verbenaceae are recovered as and Erdtman [6], Wunderlich [7] recognized six subfam- sister to Thomandersiaceae [18, 21, 22]. ilies within Lamiaceae, rejecting Lamioideae as circum- Following Cantino et al. [11], Harley et al. [1]published scribed by Briquet [5] and accepting a subfamily a global, genus-level taxonomic conspectus of Lamiaceae. Nepetoideae close to that of Erdtman [6]. Cantino and Except for the ten genera Acrymia Prain, Callicarpa L., Sanders [8] revealed that Nepetoideae sensu Erdtman [6] Cymaria Benth., Garrettia H.R. Fletch., Holocheila (Kudô) is monophyletic with several synapomorphies, whereas S. Chow, Hymenopyramis Wall. ex Griff., Ombrocharis no synapomorphy was found for Lamioideae sensu Erdt- Hand.-Mazz., Peronema Jack, Petraeovitex Oliv., and Tec- man [6]. tona L. that were treated as incertae sedis, the remaining The mint family has long been thought to have 226 genera were assigned to seven subfamilies: Ajugoi- evolved from Verbenaceae-like ancestors, and these two deae, Lamioideae, Nepetoideae, Prostantheroideae, Scutel- families were considered separate largely based upon larioideae, Symphorematoideae, and Viticoideae [1]. Since Zhao et al. BMC Biology (2021) 19:2 Page 3 of 27

the publication of this classification [1], numerous molecu- and subfamilial level, we sequenced and analyzed the lar phylogenetic studies have been carried out to explore complete plastome for 175 representative taxa from all the relationships at the subfamilial [19], tribal [23–33], or currently recognized tribes in the 12 subfamilies of generic [34–50] level. However, relationships among four Lamiaceae. The focus of this study was to (1) improve subfamilies (Nepetoideae, Tectonoideae, Premnoideae, and the resolution of the phylogenetic backbone of Lamia- Ajugoideae) remain unresolved and those among some ceae, (2) modify the tribal classification of Lamiaceae tribes were also unclear in those studies. based on our results, and (3) provide a summary of the In terms of taxon number, the most comprehensively recent phylogenetic and taxonomic progress achieved sampled phylogenetic study of Lamiaceae was conducted for each subfamily and tribe. by Li et al. [19] using an ingroup sampling of 288 species from 191 genera and employing five plastid DNA re- Results gions (matK, ndhF, rbcL, rps16, and trnL-trnF). The Characteristic of plastome features and datasets backbone of this phylogeny was comprised of 12 clades, Our sequencing generated between 13,829,468 (Sipho- all provided with high branch support, and seven of cranion flavidum Y.P. Chen & C.L. Xiang) and 81,265, which corresponded to a portion of the Viticoideae and 290 (Chloanthes coccinea Bartl.) clean reads from the 50 six of the previously recognized subfamilies of Harley newly sequenced species, with the mean base coverage et al. [1]. The other five clades consisted of previously ranging from 110× (Congea tomentosa Roxb.) to 3104× incertae sedis genera and were each provided subfamilial (Lamium amplexicaule L.) estimated by the GetOrga- rank as the Cymarioideae (including Acrymia and Cym- nelle pipeline [68]. Since we failed to assemble the aria), Peronematoideae (including Hymenopyramis, Pet- complete plastome of Callicarpa americana L., the aver- raeovitex, Peronema,andGarrettia), Premnoideae age base coverage for this species is unavailable (noted (including Premna L., Gmelina L., and Cornutia L.), Cal- as “NA” in Table 1). Statistics about the assemblies for licarpoideae (including Callicarpa), and Tectonoideae each newly sequenced species are provided in Table 1. (including Tectona)[19, 51]. All plastomes exhibit a typical quadripartite structure Despite the improved resolution in our understanding of the large single-copy (LSC, 81,341–85,891 bp) and of Lamiaceae and its subfamilies, the work by Li et al. small single-copy (SSC, 9969–20,681 bp) regions, sepa- [19] was not able to clarify relationships among Nepetoi- rated by a pair of inverted repeats (IR regions, 23,085– deae, Tectonoideae, Premnoideae, and Ajugoideae, nor 31,573 bp). The chloroplast genome maps are provided were they able to provide resolution to understand the in Additional file 1 (Fig. S1). The GC content was evenly tribal classification within some subfamilies (viz. Lamioi- distributed, and the average GC content was 38.10% deae). While recent phylogenetic analyses have greatly (Additional file 2: Table S1). All the newly sequenced improved our understanding of the major lineages and and annotated plastomes in the present study were sub- classifications of Lamioideae [52, 53], the tribal member- mitted to the National Center for Biotechnology Infor- ship of Betonica L., Colquhounia Wall., Galeopsis L., mation (NCBI) database with accession numbers Metastachydium Airy Shaw ex C.Y. Wu & H.W. Li, MT473738–MT473786 (Table 1). Paralamium Dunn., and Roylea Wall. ex Benth. remains The aligned length of the combined 79 protein-coding unclear [2, 53]. Furthermore, Xiang et al. [54] identified regions (CR) is 72,082 bp. Removal of ambiguous sites four major clades within the Ajugoideae, but did not and single-taxon insertions results in an aligned length propose a formal tribal classification. The uncertain rela- of 69,822 bp (CRM), of which 41,459 sites are constant tionships among and within these subfamilies have hin- (59.38%). The aligned regions and the excluded ambigu- dered the further study of character evolution and ous sites of the individual loci are listed in Additional diversification patterns within Lamiaceae. file 3 (Table S2), and properties of the five datasets are Next-generation sequencing (NGS) provides a signifi- summarized in Table 2. cantly larger amount of DNA sequence data than has been previously available for phylogenetic studies within angiosperms [55]. While the use of complete plastome Phylogenomic analyses sequences is not a panacea [56], it has successfully re- All analyses yielded an identical topology for the ingroup solved previously intractable phylogenetic problems at the tribal level (Fig. 1; Additional files 4, 5, 6, 7: Figs. within flowering plants at multiple taxonomic levels S2, S3, S4, S5), although the support is variable among [57–65]. Concordantly, recent phylogenomic studies different datasets. All 12 subfamilies were recovered and based on plastome sequences have provided new insight well-supported in all analyses (Fig. 1; Additional files 4, into both generic and species-level relationships within 5, 6, 7: Figs. S2, S3, S4, S5). The topology recovered by Scutellarioideae [66] and Salvia [67], respectively. In the combined dataset with the ambiguously aligned posi- order to resolve the remaining ambiguities at the tribal tions excluded (CRM) is presented as the primary tree Zhao et al. BMC Biology (2021) 19:2 Page 4 of 27

Table 1 Newly sampled species in this study (NA data unavailable) Systematic assignment Species Locality Clean reads Mean GenBank coverage accession of base (x) numbers Phrymaceae Mimulus sp. The United States Botanic 19,584,540 478 MT473772 (outgroup) Garden (USBG), United States Ajugoideae Ajugeae Caryopteris forrestii Diels Lijiang, Yunnan, China 67,295,160 485 MT473742 Ajugoideae Teucrieae Schnabelia oligophylla Kunming, Yunnan, China 67,359,376 726 MT473777 Hand.-Mazz. Ajugoideae Clerodendreae Clerodendrum japonicum Kunming, Yunnan, China 69,357,954 854 MT473745 (Thunb.) Sweet Ajugoideae Clerodendreae Clerodendrum trichotomum Huairou, Beijing, China 69,621,568 536 MT473746 Thunb. Ajugoideae Rotheceae Rotheca serrata (L.) Steane Kunming, Yunnan, China 69,698,896 328 MT473776 & Mabb. Callicarpioideae – Callicarpa americana L. Gainesville, Florida, United 69,222,992 NA -- States Callicarpioideae – Callicarpa arborea Roxb. Kunming, Yunnan, China 70,066,596 341 MT473738 Callicarpioideae – Callicarpa brevipes (Benth.) Guangzhou, Guangdong, 68,119,222 383 MT473739 Hance China Callicarpioideae – Callicarpa macrophylla Vahl Kunming, Yunnan, China 69,104,110 499 MT473740 Callicarpioideae – Callicarpa peichieniana Chun Guangzhou, Guangdong, 68,759,068 215 MT473741 & S.L. Chen ex H. Ma China & W.B. Yu Cymarioideae – Cymaria dichotoma Benth. Changjiang, Hainan, China 68,070,464 1189 MT473753 Lamioideae Paraphlomideae Paraphlomis javanica Kunming, Yunnan, China 66,797,022 239 MT473773 (Blume) Prain Lamioideae Gomphostemmateae Gomphostemma lucidum Changjiang, Hainan, China 66,781,246 274 MT473764 Wall. ex Benth. Lamioideae Gomphostemmateae Chelonopsis souliei (Bonati) Litang, Sichuan, China 67,646,436 572 MT473743 Merr. Lamioideae Colquhounieae Colquhounia coccinea Wall. Kunming, Yunnan, China 66,842,836 171 MT473749 Lamioideae Colquhounieae Colquhounia seguinii Vaniot Kunming, Yunnan, China 66,760,344 337 MT473750 Lamioideae Colquhounieae Colquhounia vestita Wall. Cuona, Xizang, China 67,753,130 192 MT473751 Lamioideae Lamieae Lamium amplexicaule L. Zuogong, Xizang, China 67,339,814 3104 MT473770 Lamioideae Synandreae Macbridea alba Chapm. The United States Botanic 20,514,794 474 MT473771 Garden (USBG), United States Lamioideae Stachydeae Galeopsis bifida Boenn. Deqin, Yunnan, China 67,442,714 500 MT473759 Nepetoideae Elsholtzieae Elsholtzia densa Benth. Shangri-La, Yunnan, China 18,273,016 888 MT473757 Nepetoideae Elsholtzieae Elsholtzia rugulosa Hemsl. Kunming, Yunnan, China 67,318,028 553 MT473758 Nepetoideae Ocimeae Siphocranion flavidum Malipo, Yunnan, China 13,829,468 436 MT473778 Y.P. Chen & C.L. Xiang Nepetoideae Ocimeae Siphocranion macranthum Nanchuan, Congqing, China 13,860,798 241 MT473779 (Hook. f.) C.Y. Wu Nepetoideae Ocimeae Hanceola exserta Y.Z. Sun Hezhou, Guangxi, China 67,557,758 203 MT473765 ex C.Y. Wu Nepetoideae Ocimeae Isodon amethystoides (Benth.) Lin'an, Zhejiang, China 25,146,824 696 MT473767 H. Hara Nepetoideae Ocimeae Isodon lophanthoides (Buch.-Ham. Kunming, Yunnan, China 40,730,966 316 MT473768 ex D. Don) H. Hara Nepetoideae Ocimeae Isodon ternifolius (D. Don) Kudô Longling, Yunnan, China 32,984,960 542 MT473769 Zhao et al. BMC Biology (2021) 19:2 Page 5 of 27

Table 1 Newly sampled species in this study (NA data unavailable) (Continued) Systematic assignment Species Locality Clean reads Mean GenBank coverage accession of base (x) numbers Nepetoideae Ocimeae Coleus xanthanthus C.Y. Wu & Mengla, Yunnan, China 25,669,120 821 MT473748 Y.C. Huang Nepetoideae Menheae Dracocephalum taliense Forrest Heqing, Yunnan, China 68,863,176 446 MT473756 Nepetoideae Menheae Clinopodium abyssinicum (Benth.) Kabarnet, Baringo, Kenya 48,657,815 833 MT473747 Kuntze Peronematoideae – Garrettia siamensis H.R. Fletcher Mengla, Yunnan, China 69,566,486 1905 MT473760 Peronematoideae – Hymenopyramis cana Craib Changjiang, Hainan, China 66,946,216 298 MT473766 Premnoideae – Premna szemaoensis C. P'ei Kunming, Yunnan, China 69,409,616 477 MT473775 Premnoideae – Premna vietnamensis Bo Li K'Bang, Gia Lai, Vietnam 80,675,070 460 MT473774 Premnoideae – Gmelina arborea Roxb. ex Sm. Mengla, Yunnan, China 67,974,942 493 MT473761 Premnoideae – Gmelina hainanensis Oliv. Kunming, Yunnan, China 67,354,640 1527 MT473762 Premnoideae – Gmelina philippensis Cham. Mengla, Yunnan, China 69,953,046 479 MT473763 Prostantheroideae Chloantheae Chloanthes coccinea Bartl. Australian National Botanic 81,265,290 598 MT473744 Gardens (ANBG), Australia Prostantheroideae Chloantheae Dasymalla teckiana (F. Muell.) Australian National Botanic 41,308,508 519 MT473754 B.J. Conn & Henwood Gardens (ANBG), Australia Prostantheroideae Chloantheae Dicrastylis parvifolia F. Muell. Australian National Botanic 81,081,410 577 MT473755 Gardens (ANBG), Australia Symphorematoideae – Congea tomentosa Roxb. Mengla, Yunnan, China 40,494,132 110 MT473752 Symphorematoideae – Sphenodesme mollis Craib Mengla, Yunnan, China 81,008,454 529 MT473780 Tectonoideae – Tectona grandis L. f. Mengla, Yunnan, China 40,169,710 514 MT473781 Viticoideae – Vitex glabrata R. Br. Mengla, Yunnan, China 70,126,282 722 MT473782 Viticoideae – Vitex negundo var. cannabifolia Kunming, Yunnan, China 67,083,468 1387 MT473783 (Siebold & Zucc.) Hand.-Mazz. Viticoideae – Vitex quinata (Lour.) F.N. Williams Mengla, Yunnan, China 69,282,366 828 MT473784 Viticoideae – Vitex tripinnata (Lour.) Merr. Guangzhou, Guangdong, 67,065,514 1404 MT473785 China Viticoideae – Vitex yunnanensis W.W. Sm. Luquan, Yunnan, China 70,217,642 395 MT473786

(Fig. 1) for the following discussion of phylogenetic Within Nepetoideae (100%, 1.00), the monophyly of relationships. Elsholtzieae, Ocimeae, and Mentheae was robustly sup- Within Lamiaceae, two primary clades were recovered ported in all analyses (100%, 1.00). However, relation- and subdivided as 12 clades corresponding to the 12 sub- ships among the three tribes varied among different families (Fig. 1), with each subfamily being monophyletic datasets. Most of the datasets (CRM, CR, CR3, dePCS) (excepting Cymarioideae, which was represented by only supported Elsholtzieae as sister to Ocimeae (Fig. 1, 86%, one species). The first clade comprised the Prostantheroi- 1.00; Additional files 4, 5: Figs. S2, S3; Additional file 7: deae and Callicarpoideae (i.e., Calliprostantherina sensu Li Fig. S5), while in the phylogeny based on dataset CR12, et al. [19]), both with strong support (MLBS = 100%, Elsholtzieae were weakly supported as sister to BIPP = 1.00; Fig. 1;Additionalfiles4, 5, 6, 7:Figs.S2,S3, Mentheae (Additional file 6: Fig. S4, 45%, 0.66). S4, S5, and all support values follow this order hereafter). In tribe Elsholtzieae, the genus Elsholtzia Willd. was The two tribes of Prostantheroideae, Chloantheae and recovered as sister to Collinsonia L. and Perilla L., and Westringieae, were each recovered as monophyletic and the sister relationships received maximal support in all sister taxa with strong support (100%, 1.00). The second analyses (Fig. 1; Additional files 4, 5, 6, 7: Figs. S2, S3, clade of Lamiaceae consisted of Nepetoideae, Symphore- S4, S5). Representatives of all seven subtribes of Oci- matoideae, Viticoideae, Tectonoideae, Premnoideae, Aju- meae formed a well-resolved clade, with subtribe Sipho- goideae, Peronematoideae, Scutellarioideae, Cymarioideae, cranioninae (Siphocranion spp.) diverging first, followed and Lamioideae (Fig. 1;Additionalfiles4, 5, 6, 7: Figs. S2, by subsequent bifurcations for subtribes Lavandulinae S3, S4, S5). (Lavandula spp.), Hanceolinae (Hanceola exserta Y.Z. Zhao et al. BMC Biology (2021) 19:2 Page 6 of 27

Table 2 Data characteristics with models selected for each dataset used for phylogenetic study in the present study Dataset CRM CR CR12 CR3 dePCS GC content 38.3% 38.3% 40.2% 34.5% 30.8% Alignment sites (bp) 69,822 72,082 48,069 24,013 72,082 Constant sites (bp) 41,459 43,415 31,083 12,331 50,977 Parsimony-informative sites (bp) 29,945 20,185 11,561 8,624 14,473 Variable sites (bp) 28,363 28,667 16,986 11,682 21,105 Missing data 4.31% 4.31% 4.31% 4.31% 4.31% Best-fit model GTR+I+G GTR+I+G GTR+I+G GTR+I+G GTR+G

Sun ex C.Y. Wu), Isodoninae (Isodon spp.), Hyptidinae were only represented by a limited number of species (Mesosphaerum suaveolens (L.) Kuntze), Ociminae (Oci- (e.g., Lamieae, Leucadeae, and Leonureae). mum spp.), and Plectranthinae (Coleus spp.). Relation- ships within tribe Mentheae were also well resolved Discussion (100%, 1.00), with subtribe Salviinae recovered as sister It has been more than 20 years since the first attempt to the remaining four subtribes, Prunellinae, Lycopinae, was made to employ molecular data as evidence to infer Menthinae, and Nepetinae. a phylogenetic tree for Lamiaceae, which made use of Along the backbone of the tree, subsequent to the the rbcL region of the chloroplast genome [15]. Subse- branching of the Nepetoideae, Symphorematoideae quently, various phylogenetic analyses have greatly con- (100%, 1.00) and Viticoideae (100%, 1.00) formed a clade tributed to our understanding of the circumscription, (i.e., Viticisymphorina sensu Li et al. [19]), which was classification, and phylogeny of this family, progressively followed by subsequent bifurcation supporting clades of improving the resolution of relationships [15, 19, 25, the Tectonoideae (100%, 1.00), Premnoideae (100%, 27–31, 44, 46, 52–54, 69]. This study, based on coding 1.00), and then Ajugoideae, respectively (Fig. 1, 100%, plastome sequences, provides the most comprehensive 1.00). Ajugoideae (100%, 1.00) were divided into four phylogeny of Lamiaceae at the tribal level to date. With subclades that corresponded with the structure of tribal increased taxon sampling and a vastly expanded DNA classification: each tribe was recovered as monophyletic dataset, the results of our plastid phylogeny significantly and provided with high branch support (100%, 1.00). clarify the remaining ambiguities for all relationships Within the Ajugoideae, Rotheceae were recovered as sis- among subfamilies and provide better support for all ter to the Teucrieae, Clerodendreae, and Ajugeae. nodes in the phylogenetic tree at the subfamilial level. The sister clade of Ajugoideae was comprised of Pero- In our phylogenetic analyses, 12 subfamilies are recov- nematoideae, Scutellarioideae, Cymarioideae, and ered and well-supported as monophyletic (Fig. 1; Add- Lamioideae (i.e., the phylogenetically defined Perola- itional files 4, 5, 6, 7: Figs. S2, S3, S4, S5). Our results miina in Li et al. [19]). Monophyly of Ajugoideae plus correspond with the most recent phylogenetic study Perolamiina was supported in all analyses with moderate using five cpDNA regions [19] and have resolved the support values (Fig. 1, 71%, 0.98; Additional files 4, 5, 6, placement of the Nepetoideae, Premnoideae, and Aju- 7: Figs. S2, S3, S4, S5), and Peronematoideae were recov- goideae which were previously unknown. Nepetoideae, ered as monophyletic (100%, 1.00) and sister to Scutel- the largest subfamily of Lamiaceae, is sister to a grade of larioideae + Cymarioideae + Lamioideae (i.e., lineages comprising the Symphorematoideae, Viticoi- Scutelamiina sensu Li et al. [19]). Within Scutellarioi- deae, Tectonoideae, Premnoideae, Ajugoideae, Perone- deae, four out of five genera were included for analyses matoideae, Scutellarioideae, Cymarioideae, and and the monotypic genus Wenchengia C.Y. Wu & S. Lamioideae (Fig. 1). However, our results differ some- Chow (100%, 1.00) is sister to the remaining three what from those of the Mint Evolutionary Genomics genera (100%, 1.00). The sister clade of Scutellarioideae Consortium [20], which used 520 single-copy nuclear consisted of Cymarioideae and Lamioideae (100%, 1.00). genes from 48 Lamiaceae species representing 11 of 12 Within Lamioideae, Pogostemoneae were the earliest di- subfamilies. Their results of the first-diverging lineages verging lineage, followed by the Gomphostemmateae, were consistent with ours and only differ within the Colquhounieae, Synandreae, Betoniceae, Galeopseae, clade of Premnoideae, Ajugoideae, Peronematoideae, Stachydeae, Paraphlomideae, Phlomideae, Leonureae, Scutellarioideae, Cymarioideae, and Lamioideae, where Marrubieae, Leucadeae, and Lamieae; consistent with most of the relationships in their tree were weakly sup- previously published studies [52, 53], most tribes re- ported. Furthermore, taxon sampling was sparse in their ceived maximal support values, although some tribes study, and it is possible that additional taxon sampling Zhao et al. BMC Biology (2021) 19:2 Page 7 of 27

Fig. 1. Maximum likelihood phylogeny of Lamiaceae based on combined 79 plastid coding regions dataset, with ambiguously aligned sites excluded. Maximum likelihood bootstrap support (MLBS) and Bayesian inference posterior probability (BIPP) are shown above and below the branches, respectively. Bold horizontal lines indicate clades with BIPP = 1.00) and MLBS = 100%. A “–” indicates MLBS values < 50% and BIPP < 0.8. Subfamilies and tribes recognized by Li et al. [19] and Li and Olmstead [51] are indicated by gray boxes, while new tribes proposed in this study were marked in red font

could alter the subfamilial relationships that their ana- lyses recovered. Relationships within Lamioideae are also relatively similar with previous broad-scale studies [52, 53], but in- ternal support values from our study are generally higher. Within Lamioideae, five genera (Betonica, Colqu- hounia, Galeopsis, Metastachydium,andRoylea) have not previously been assigned tribal status [2, 52, 53]. In addition, the phylogenetic position of Paralamium re- mains unclear [2, 53], since the genus has not been in- cluded in any published molecular phylogenetic study. We included three of these genera (Betonica, Colquhou- nia, and Galeopsis) in our study. Colquhounia is recovered as sister (Fig. 1, 100%, 1.00) to the clade of Synandreae, Betoniceae, Galeopseae, Sta- chydeae, Paraphlomideae, Phlomideae, Leonureae, Mar- rubieae, Leucadeae, and Lamieae. The morphological distinctiveness and well-supported phylogenetic position of Colquhounia substantiates tribal recognition within Lamioideae as tribe Colquhounieae (see “Taxonomic treatment”). Corroborating previous phylogenetic studies [52, 53], our chloroplast phylogeny demonstrates that Galeopsis and Betonica form a clade (Fig. 1, 64%, 0.98) that is sis- ter to the Stachydeae (100%, 1.00). This clade in turn is recovered as sister to a clade of Paraphlomideae, Phlo- mideae, Leonureae, Marrubieae, Lamieae, and Leuca- deae. Using cpDNA markers, Scheen et al. [52] and Bendiksby et al. [53] found this same structure, and our unpublished data based on chloroplast DNA markers (M. Bendiksby and Y. Salmaki, in prep.) also suggests these two genera occupy different positions within Lamioideae. In contrast, analyses using the low-copy nu- clear pentatricopeptide repeat (PPR) region recovered Galeopsis as sister to tribe Synandreae rather than sister to Betonica, albeit this was provided with low support [69]. With the available evidence (see “Discussion”), the phylogeny supports that Betonica and Galeopsis are dis- tinct from other tribes. As suggested by Li and Olmstead [51], “for the benefit of those who need a complete, rank-based classification of Lamiaceae to arrange genera and species in checklists”, a new monotypic tribe (i.e., Betoniceae) is established here and the tribe Galeopseae (also monotypic) is resurrected, to accommodate the Zhao et al. BMC Biology (2021) 19:2 Page 8 of 27

systematic positions of these two genera within Lamioi- [73]. This distribution includes a large number of species deae. The tribal placement of the remaining three gen- adapted to extreme arid habitats, with genera such as New- era, Paralamium, Roylea, and Metastachydium, is still castelia F. Muell. and Dicrastylis Drumm. ex Harv. occupy- uncertain. ing sandy deserts of the central inland [74]. Within Ajugoideae, we recover the same relationships A remarkable diversity in floral morphology is dis- as reported by Xiang et al. [54], who sampled 51 taxa played across Chloantheae, with corollas ranging from representing 22 of the 23 genera of the subfamily and 5-merous and zygomorphic (e.g., Chloanthes R. Br. and identified four main clades. All clades are recovered as Dasymalla Endl.) to 5–8(–10)-merous and actino- monophyletic and receive better resolution (Fig. 1). Al- morphic (e.g., Dicrastylis). All species are distinguished though Xiang et al. [54] improved our understanding of (particularly from the sister tribe Westringieae) by an relationships within Ajugoideae, a tribal classification unlobed ovary, which develops into a 1 (–2) seeded dry scheme for the subfamily has been needed. Corroborating indehiscent fruit [1], and a distinctive indumentum of previous studies [54], we propose a formal tribal classifica- complex dendritic trichomes (typically tomentose) cov- tion for subfamily Ajugoideae, including the new tribe ering branches, leaves, and flowers (except four species Rotheceae (see “Taxonomic treatment”). in the Westringieae). The advances in our knowledge reported in the results Many taxonomic changes have been made for above cement a foundation in our understanding of rela- Chloantheae and its constituents. Since the description tionships within Lamiaceae. In order to provide a clearer of Chloanthes and Pityrodia R. Br. [75], most genera picture in light of these results and to consolidate the were shuffled between different tribes of Verbenaceae numerous advances made in the systematics of Lamia- [76, 77]. Most were allocated within the tribe ceae since Harley et al. [1], the following sections pro- Chloantheae (Verbenaceae) by Bentham [4]. This treat- vide a detailed discussion and commentary for each ment was followed later by Hutchinson’s recognition as subfamily and tribe. family Chloanthaceae [78], which was accepted by some authors [74, 79–83], but not all [84, 85]. Subfamily Prostantheroideae Luerss. Phylogenetic analysis of morphological [9] and mo- Prostantheroideae consist of approximately 315 species lecular data [71] indicated that Chloantheae is sister to allocated to two tribes: Chloantheae and Westringieae. Westringieae within Lamiaceae, which is supported here They are distinguished from all other subfamilies by hav- (Fig. 1). The contemporary understanding of generic re- ing a prominent albuminous seed [4, 8]. While multiple lationships within the tribe was informed by the compre- cell layers can be found in the endosperm in other sub- hensively sampled molecular phylogeny of Conn et al. families [70] (therefore technically albuminous), the [24], which found that Pityrodia was not monophyletic, endosperm never develops to a size that can be easily precipitating the description of Muniria N. Streiber & seen [1]. B.J. Conn and restoration of Dasymalla and Quoya Gau- Although confined to Australia, Prostantheroideae are dich. [73]. Another new genus, Apatelantha, was re- widely distributed throughout most of the continent, in cently described to accommodate a clade identified by both temperate and tropical climates. Within this ex- Conn et al. [24] composed of individuals formerly panse, the habitats they occupy range from riparian assigned to Lachnostachys Hook., Newcastelia,andPhy- zones of cool temperate rainforest to crests of shifting sopsis Turcz. [86]. Although our study only samples sand dunes in the central arid region. three taxa in Chloantheae, as in previous studies [73], it Prostantheroideae are sister to Callicarpoideae (i.e., Cal- supports the close relationship between Dasymalla and liprostantherina sensu Li et al. [19]). This relationship was Chloanthes relative to Dicrastylis (Fig. 1; Additional files first discovered by Olmstead et al. [71], then consistently 4, 5, 6, 7: Figs. S2, S3, S4, S5). supported by subsequent molecular phylogenetic studies [18–20, 31, 72] as well as our own (Fig. 1; Additional files Tribe Westringieae Bartl. 4, 5, 6, 7: Figs. S2, S3, S4, S5). Together, both Prostanther- Westringieae consist of five genera and over ca. 210 spe- oideae and Callicarpoideae form a sister clade to the cies of subshrubs, shrubs, and small trees distributed remaining Lamiaceae (Fig. 1)[18–20, 31, 72]. In addition across Australia [1]. Frequently found restricted to ex- to having albuminous seeds, Prostantheroideae are distin- posed and rocky or well-drained places, members of the guished from Callicarpoideae by their dry fruits (vs. fleshy tribe are distributed throughout habitats within which fruits). these places occur, from rainforests to ranges of the Australian arid inland. Tribe Chloantheae Benth. & Hook. f Flowers are 5-merous and weakly to strongly zygo- Chloantheae consist of 13 genera and ca. 100 species of morphic, similar to bird or insect pollination syndromes shrubs (or subshrubs) distributed across mainland Australia typically found in other Lamiaceae [87–89]. The tribe Zhao et al. BMC Biology (2021) 19:2 Page 9 of 27

can be distinguished from Chloantheae by a four-lobed Subfamily Nepetoideae (Dumort.) Luerss. ovary, which develops into four nutlets [1]. The variation Nepetoideae are the most species-rich subfamily within in anther morphology (e.g., outgrowth of the antheridial Lamiaceae, with about 3400 species divided into three connective of Prostanthera Labill.) combined with re- tribes, Elsholtzieae, Mentheae, and Ocimeae [1]. Nepe- ductions in fertility (reduction of abaxial stamens to sta- toideae are native to every continent except Antarctica minodes in Westringia Sm.) in this tribe distinguishes it and are found in each of the seven global regions of high from Chloantheae (which typically has four bithecate an- Lamiaceae diversity [1, 93]. Although only clarified when thers) and assists with informing the contemporary gen- comparative pollen analyses were established [6, 8], eric delimitation in the tribe [1, 90]. Nepetoideae are now considered among the most clearly Tribal recognition of Westringieae and its generic con- defined subfamilies of Lamiaceae and have consistently stituency was first described by Bentham [91]. The been supported as monophyletic in molecular analyses monophyly of this tribe, in addition to its sister relation- [15, 19, 31, 44, 94, 95]. Nepetoideae contain nearly all ship to Chloantheae, has been substantiated by numer- the aromatic species within Lamiaceae and are charac- ous phylogenetic analyses [9, 19, 71] including our own terized by hexacolpate, trinucleate pollen [6, 8], an (Fig. 1). Further investigation into generic relationships investing embryo [96], and the presence of rosmarinic has shown that Hemiandra R. Br., Hemigenia R. Br., acid [1]. Additionally, mucilaginous nutlets are only Microcorys R. Br., and Westringia are closely related to known to occur in the Nepetoideae within Lamiaceae each other with respect to Prostanthera [87, 90, 92], al- and occur in all three tribes [97]. Thus, mucilaginous though the relationship between them still needs to be nutlets may also represent a synapomorphy within resolved by more comprehensively sampled phylogenetic Nepetoideae. studies. The tribal assignment for groups now in Nepetoideae has been controversial [4, 5, 7] and was summarized by Cantino [10]. Results from morphological and molecular Subfamily Callicarpoideae Bo Li & R.G. Olmstead studies [9, 10, 95] led to a fundamentally new tribal clas- This recently described subfamily consists only of the sification for Nepetoideae proposed by Cantino et al. genus Callicarpa which contains ca. 170 species of [11]. They recognized the four tribes Elsholtzieae, Oci- small trees or shrubs primarily distributed in tropical meae, Lavanduleae, and Mentheae, with the latter con- to temperate Asia, tropical and subtropical America, taining the largest number of changes in Australia, and some Pacific Islands [19, 51]. Callicar- circumscription. Harley et al. [1] basically adopted this poideae differs from other subfamilies by having a treatment of Cantino et al. [11], with the exception of peltate or capitate stigma and a drupaceous fruit with subsuming Lavanduleae within Ocimeae. Although the four stony pyrenes [51]. Furthermore, Callicarpoideae three tribes of Harley et al. [1] are well-supported in possess actinomorphic flowers which are unusual both previous studies [16, 23, 27, 31, 98] and our ana- within Lamiaceae (generally zygomorphic). The group lyses (Fig. 1; Additional files 4, 5, 6, 7: Figs. S2, S3, S4, is remarkably morphologically homogeneous given its S5), relationships among the three tribes remain murky. broad geographical distribution, although there is Previous studies have either found (1) Ocimeae to be sis- variation in the number of flower parts and stamen ter to the Mentheae-Elsholtzieae clade [95], or (2) structure among different species within Mentheae to be sister to the Ocimeae-Elsholtzieae clade Callicarpoideae. [16, 23, 27, 98], or (3) Elsholtzieae to be sister to the Callicarpa was historically placed in Verbenaceae and Mentheae-Ocimeae clade [31]. Our results reveal that treated as a member of tribe Callicarpeae in subfamily Elsholtzieae is sister to Ocimeae in most of the analyses Viticoideae [5]. It was first transferred to Lamiaceae (CRM, CR, CR3, dePCS) (Fig. 1; Additional files 4, 5, 7: based on a cladistic analysis of morphological, anatom- Figs. S2, S3, S5), but is weakly supported as sister to ical, and palynological characters [9, 10] and later con- Mentheae by the dataset CR12 (Additional file 6: Fig. firmed by molecular study [19]. Because only one or few S4). Since none of the abovementioned relationships are representatives of the genus were included, different strongly supported, nor a broad sampling within all phylogenetic analyses resolved Callicarpa in different three tribes are included in these studies, further studies positions within Lamiaceae [19, 31, 52, 53, 71]. are still needed to resolve the relationships among the The sister relationship between Callicarpa and Pros- three tribes. tantheroideae was first discovered by Olmstead et al. [71] and confirmed by subsequent studies [18–20, 31, Tribe Elsholtzieae (Burnett) R.W. Sanders & P.D. Cantino 72]. In our analyses, they form a well-supported clade, Elsholtzieae are the smallest tribe of Nepetoideae, com- which is sister to the remaining Lamiaceae (Fig. 1; Add- prising eight genera and ca. 70 species mostly distributed itional files 4, 5, 6, 7: Figs. S2, S3, S4, S5). across East and Southeast Asia. Collinsonia, which is Zhao et al. BMC Biology (2021) 19:2 Page 10 of 27

restricted to eastern North America, is the sole New sedis within Ocimeae, while Harley et al. [105] later World member of this tribe [1, 98]. Species of Elsholt- established subtribe Hanceolinae to accommodate them. zieae share divergent stamens, a weakly 2-lipped corolla, Paton et al. [23] carried out the first molecular phylo- and an asymmetric disc with an elongate anterior lobe, genetic analyses of Ocimeae and revealed that the genus but it is unclear whether these features are apomorphic Lavandula L. was sister to the remaining Ocimeae and [1, 31]. thus subtribe Lavandulinae was recognized within The tribe was formally validated by Sanders and Can- Ocimeae [23]. However, the two genera Hanceola and tino [99] and consisted of six genera in the classification Siphocranion were not included in their analysis. The of Cantino et al. [11]: Collinsonia, Elsholtzia, Keiskea phylogenetic relationships within Ocimeae were fur- Miq., Mosla (Benth.) Buch.-Ham. ex Maxim., Perilla, ther elucidated based on more comprehensive sam- and Perillula Maxim. In the molecular phylogenetic pling by Zhong et al. [106], who demonstrated that study of Nepetoideae by Wagstaff et al. [95], Elsholtzieae Siphocranion, Hanceola,andIsodon each formed a was represented by Elsholtzia, Collinsonia, and Perilla distinct lineage within Ocimeae. The subtribes Sipho- and formed a well-supported clade. Based on a sampling cranioninae and Isodoninae were thus described to of all genera of Elsholtzieae using two nrDNA and four accommodate Siphocranion and Isodon, respectively, cpDNA markers, the results by Chen et al. [31] con- while subtribe Hanceolinae only includes Hanceola firmed that the previously incertae sedis genus Ombro- [106]. charis is a member of the tribe and sister to Perillula. Recently, Chen et al. [107] reported a new species of Contemporaneously, based on results from molecular Siphocranion, and in their molecular phylogenetic ana- phylogenetic analyses [31] and karyological studies [100], lyses based on six cpDNA markers, Siphocranioninae is Mayta-Anco et al. [101] established a new genus, shown to be sister to the remaining subtribes, with Vuhuangia Solomon Raju, Molinari & Mayta, to accom- Lavandulinae further supported as the sister group of modate Elsholtzia flava (Benth.) Benth. and E. penduli- the clade including Hanceolinae, the Isodoninae- flora W.W. Sm. However, Li et al. [98], apparently Hyptidinae clade, and the Plectranthinae-Ociminae unaware of Vuhuangia, demonstrated that Elsholtzia clade. Our phylogenomic analyses largely confirm the re- was not monophyletic and outlined E. flava and E. pen- sults of Chen et al. [107], with the exception that Isodo- duliflora should be separated from Elsholtzia as a dis- ninae is resolved as sister to the Hyptidinae-Ociminae- tinct genus. Plectranthinae clade (Fig. 1; Additional files 4, 5, 6, 7: Biogeographic analysis of an expanded sample of Figs. S2, S3, S4, S5). Elsholtzieae showed that the tribe originated in East Asia and then dispersed to Southeast Asia and North Amer- Tribe Mentheae Dumort. ica; the uplifts of the Qinghai-Tibetan Plateau and cli- Mentheae are characterized by stamens divergent or as- mate changes from Middle Miocene onwards may have cending (not declinate), a distinctly 2-lipped corolla promoted the species diversification of Elsholtzieae [98]. (rarely weakly so), symmetric disc (if asymmetric and an- terior lobe elongate, then corolla distinctly 2-lipped), Tribe Ocimeae Dumort. and nutlets with an areolate abscission scar. Some of the Ocimeae are characterized by declinate stamens lying along most widely known medicinal and culinary plants are the anterior lip of the corolla and synthecous anthers [1, found within this group: mint, oregano, sage, savory, and 102]. As currently circumscribed, a total of 43 genera and thyme. Mentheae comprise both the largest number of over 1200 species are included in Ocimeae, distributed genera and species of any tribe within Nepetoideae and mainly in the tropics and subtropics [1, 103, 104]. Major Lamiaceae. Many of the plants in this group are of eco- centers of diversity include tropical Africa and Madagascar, nomic and ecological importance and thus have com- China and Malaysia, and South America [1, 103]. monly attracted the attention of scientists. This has In early classifications of Lamiaceae [4, 5], Ocimeae resulted in fundamentally differing taxonomic ap- were recognized as subfamily Ocimoideae. Based on an proaches at all taxonomic ranks, making it difficult to expansive morphological cladistic analysis, Cantino [9, provide accurate numbers for genera (about 60) or spe- 10] reduced Ocimoideae to tribe Ocimeae within sub- cies (at least 2000). family Nepetoideae sensu Cantino et al. [11]. Ocimeae Due to the abovementioned fluidity regarding circum- was further divided into three subtribes: Hyptidinae, scription within Mentheae, the classification of Harley et al. Plectranthinae, and Ociminae [11]. Because Isodon [1] is regarded as the starting point for a modern subtribal (Schrad. ex Benth.) Spach, Hanceola Kudô, and Sipho- classification. There, three subtribes were recognized, cranion Kudô are very different from other Ocimeae in Menthinae, Nepetinae, and Salviinae, along with two genera terms of nutlet, inflorescence, and calyx morphology, Pa- of uncertain placement (Heterolamium C.Y. Wu and Me- ton and Ryding [102] treated the three genera as incertae lissa L.). Since the treatment of Harley et al. [1], Zhao et al. BMC Biology (2021) 19:2 Page 11 of 27

relationships within Menthinae have been greatly clarified Ajugoideae and Scutellarioideae. Furthermore, the type based on molecular phylogenetic studies [25, 27, 108–110]. genus of Viticoideae, Vitex, has expanded to include Drew and Sytsma [27] accommodated Cleonia L., Hormi- Paravitex H.R. Fletcher, Petitia Jacq., Tsoongia Merr., num L., and Prunella L. in Prunellinae and erected a new and Viticipremna H.J. Lam based on molecular studies subtribe, Lycopinae, for the enigmatic genus Lycopus L. (a [19, 42]. Even though only three genera remain in Viti- tribe Lycopeae was previously proposed [111]). Neoeplingia coideae as currently circumscribed, the intergeneric re- Ramamoorthy, Hiriart & Medrano along with Melissa were lationships are still questionable, with the positions of transferred to Salviinae [27]whileHyssopus L. and the pre- Teijsmanniodendron and Pseudocarpidium poorly re- viously unplaced Heterolamium were included in Nepetinae solved [19]. As mentioned above, the sister relationship based on morphological [112] and molecular results [27, between Viticoideae and Symphorematoideae is firmly 113]. The currently accepted number of subtribes is thus supported, and the two subfamilies share several ana- five.Thisisalsowell-supported by our analyses, where Sal- tomical traits [19]. Morphologically, species of Viticoi- viinae is sister to the other four subtribes; among the deae can be easily recognized by the palmately remaining subtribes, Nepetinae and Menthinae are sister compound leaves and dry or fleshy drupes or groups, with Prunellinae and Lycopinae as successive sister schizocarps. groups to Nepetinae and Menthinae (Fig. 1; Additional files 4, 5, 6, 7:Figs.S2,S3,S4,S5). Subfamily Tectonoideae Bo Li & R.G. Olmstead Tectonoideae comprise only the three species of Tec- Subfamily Symphorematoideae Briq. tona. They are large trees native to tropical Asia from Symphorematoideae contain about 21 species in three India to Southeast Asia, but are widely cultivated and genera of woody climbers, Congea Roxb., Sphenodesme naturalized in Africa, Central and South America, and Jack, and Symphorema Roxb., and occur mainly in trop- the Caribbean [51]. ical regions of Asia. Symphorematoideae are character- Tectona was originally placed in tribe Tectoneae of ized by having capitate cymes surrounded by bracteoles Viticoideae [5], but was revealed to be sister to a large which are often conspicuous, colorful, and accrescent, clade comprising Lamioideae, Cymarioideae, Scutellar- and incompletely 2-locular ovaries [19]. ioideae, Peronematoideae, Ajugoideae, and Premnoideae Historically, Symphorematoideae has been treated as a [19]. The relationship is also confirmed by our analyses separate family with the same circumscription [114, 115] (Fig. 1; Additional files 4, 5, 6, 7: Figs. S2, S3, S4, S5). or (more commonly) as part of Verbenaceae [5, 116]. It However, Tectona was recovered as sister to a larger was first found to be related to Lamiaceae in the mo- clade including the aforementioned subfamilies (Cymar- lecular era [15, 16], and then transferred to Lamiaceae ioideae not sampled) as well as Symphorematoideae and and treated as a subfamily [1, 117]. Li et al. [19] were Viticoideae in an analysis using low-copy nuclear markers the first to include all three genera of Symphorematoi- [20]. Regardless of phylogenetic position, Tectonoideae deae in a comprehensive phylogenetic analysis of Lamia- represents a genetically isolated clade in Lamiaceae and ceae based on chloroplast sequences, and has a series of distinct morphological traits [19, 51]. Symphorematoideae was found to be monophyletic and sister to Viticoideae. Such a sister relationship was fur- Subfamily Premnoideae Bo Li, R.G. Olmstead & P.D. ther recovered in phylogenetic analyses based on nuclear Cantino genes [20] and confirmed in our phylogenomic analyses Premnoideae were recently established to include three using plastome sequences (Fig. 1; Additional files 4, 5, 6, former viticoid genera (Sensu Harley et al. [1]): Cornu- 7: Figs. S2, S3, S4, S5). tia, Gmelina, and Premna [19], with the total species number estimated at about 150 (B. Li, pers. comm.). Subfamily Viticoideae Briq. Nearly all species of this subfamily are woody shrubs, Viticoideae currently include ca. 280 species in three trees, or climbers, occurring mainly in Old World trop- genera: Vitex (250 spp.), Teijsmanniodendron Koord. (23 ical to subtropical regions (Gmelina and Premna) and spp.), and Pseudocarpidium Millsp. (9 spp.). These gen- the New World Tropics (Cornutia)[19]. era are distributed predominantly in the Tropics with a With the current circumscription, Premnoideae are few species of Vitex occurring in temperate regions of well-supported in our phylogenomic trees (Fig. 1; the Northern Hemisphere [19]. Additional files 4, 5, 6, 7: Figs. S2, S3, S4, S5). However, Viticoideae as defined by Briquet [5] were a heteroge- in a phylogeny of Lamiaceae based on nuclear genes, neous group whose circumscription has shrunk dramat- Cornutia was not recovered in Premnoideae but was sis- ically. Segregated from traditional Viticoideae are three ter to the Lamioideae-Ajugoideae-Peronematoideae-Scu- subfamilies, Callicarpoideae, Premnoideae, and Tecto- tellarioideae clade [20, 72]. In the analyses of Li et al. noideae in the present classification, and part of [19], the relationships among Premnoideae, Ajugoideae, Zhao et al. BMC Biology (2021) 19:2 Page 12 of 27

and Lamioideae-Cymarioideae-Scutellarioideae-Perone- Rotheca, the largest genus in this tribe, was resurrected matoideae were not well resolved, but in our phyloge- by Steane and Mabberley [123] to maintain the mono- nomic analyses, Premnoideae are strongly supported to phyly of the genus Clerodendrum [35]. In the present be sister to the clade comprising Lamioideae, Cymarioi- study, we demonstrate Rotheca to be sister to all other deae, Scutellarioideae, Peronematoideae, and Ajugoideae members of the subfamily, as reported by Yuan et al. (Fig. 1; Additional files 4, 5, 6, 7: Figs. S2, S3, S4, S5). [124]. Although only Rotheca was sampled here, a close relationship to the other three genera has been demon- Subfamily Ajugoideae Kostel. strated previously [54]. Steane et al. [36] found Karomia Ajugoideae are the third-largest subfamily within Lamia- to be sister to Rotheca based on ndhF sequences, and ceae and contain about 770 species in 23 genera [19, 48, this relationship was corroborated by Li et al. [19] based 54, 118, 119] distributed worldwide but most common on five cpDNA markers. Xiang et al. [54] found that in tropical regions [1]. A possible synapomorphy of Aju- Karomia, Discretitheca, Glossocarya, and Rotheca goideae may be pollen with branched to granular colu- formed a clade, but with moderate support. Discretitheca mellae [9]. and Glossocarya were only first included in molecular Briquet [5] first elevated tribe Ajugeae sensu Bentham phylogenetic analyses [54], and detailed morphological [4] to subfamilial rank, which was followed by most sub- studies as well as molecular phylogenetic studies for sequent treatments [1, 7, 116, 120]. Circumscription of these two genera are scarce and more studies are Ajugoideae, however, has changed considerably. The needed. As with Discretitheca and Glossocarya, only one recognition of some subfamilies (i.e., Teucrioideae and species of Karomia (K. speciosa (Hutch. & Corbishley) R. Caryopteridoideae) that include many traditionally ver- Fern.) has been included in previous molecular phylo- benaceous genera (e.g., Caryopteris Bunge, Cleroden- genetic analyses [36, 54], although DNA sequences of drum L., Schnabelia Hand.-Mazz., and Teucrium L.) was two species have been reported (the additional species is untenable. These genera were later transferred to Aju- K. tettensis (Klotzsch) R. Fern. which was used mainly goideae based on molecular phylogenetic [15, 16] and for ecological analyses [125]). Overall, the systematic re- morphological evidence [121]. lationships within this tribe await to be fully clarified. A recent phylogenetic study that sampled 22 out of the 23 genera of Ajugoideae and used four cpDNA Tribe Teucrieae Dumort. markers (matK, rbcL, trnL-trnF, and rps16) strongly sup- Teucrieae consist of ca. 260 species in three genera, Teu- ported the monophyly of Ajugoideae and identified four crium (ca. 250 spp.), Schnabelia (5 spp.), and Rubiteucris major clades [54]. Relationships among these clades are Kudô (2 spp.). The latter two genera are endemic to East consistent with the results in our study. Asia, while Teucrium has a subcosmopolitan distribu- Currently, no tribal classification has been assigned for tion. A possible synapomorphy of the tribe is the conflu- Ajugoideae. Although some old tribal names have been ence of anther thecae at anthesis, a feature that also proposed [5, 91, 122], the circumscription of Lamiaceae characterizes Ajugeae, where it may have arisen at that time was much narrower compared to our independently. current understanding, and many genera now placed Teucrium is the largest genus in this tribe. A previ- within Ajugoideae (e.g., Caryopteris, Clerodendrum, ous phylogenetic study [48] suggested the inclusion Rotheca, Schnabelia, Volkameria L.) were previ- of Oncinocalyx F. Muell., Spartothamnella Briq., and ously treated as members of Verbenaceae. Based on re- Teucridium Hook.f. in Teucrium, and this treatment sults from both the present and previous studies [19, was confirmed by Xiang et al. [54]. Although both 54], we suggest that the four clades be recognized as Rubiteucris and Schnabelia are small genera, the tax- tribes Ajugeae, Clerodendreae, Teucrieae, and Rothe- onomy and systematic relationships of Rubiteucris ceae, with the last proposed here as a new tribe (see and Schnabelia were not sufficiently resolved until “Taxonomic treatment” below). recent molecular phylogenetic studies based on a broad sampling [48, 54]. Here, the monophyly of Tribe Rotheceae Teucrieae is strongly supported (Fig. 1; Additional Rotheceae are established as a new tribe (see “Taxo- files 4, 5, 6, 7:Figs.S2,S3,S4,S5). nomic treatment” below) comprising four genera: Rotheca (60 spp.), Glossocarya Wall. ex Griff. (13 spp.), Tribe Ajugeae Benth. Discretitheca P.D. Cantino (1 sp.), and Karomia Dop. (9 Ajugeae contain 79 species in six genera: Ajuga L. (ca. spp.). The tribe is disjunctly distributed from Australia 50 spp.), Amethystea L. (1 sp.), Caryopteris (7 spp.), (Queensland) and tropical southern Asia to southern Af- Pseudocaryopteris (Briq.) P.D. Cantino (3 spp.), Trichos- rica. No non-molecular synapomorphy has been tema Gronov. (17 spp.), and Tripora P.D. Cantino (1 found for this tribe. sp.). Ajuga is distributed primarily in Eurasia, Zhao et al. BMC Biology (2021) 19:2 Page 13 of 27

Amethystea is widespread in temperate Asia [1], Trichos- (Tribe Rotheceae), where it apparently evolved inde- tema is restricted to North America [126], and the pendently. The character polarity is not entirely clear be- remaining three genera are endemic to East Asia. A pos- cause Premnoideae also have drupaceous fruits. sible synapomorphy is the confluence of the anther the- However, the fruits of Premnoideae contain a single cae at anthesis (with a reversal in Caryopteris), a feature four-seeded pyrene instead of four one-seeded ones. The that also characterizes Teucrieae and may have arisen other closely related groups (subfamilies Peronematoi- independently in the two tribes. In most other species of deae, Scutellarioideae, Cymarioideae, and Lamioideae) Ajugoideae and in most of the closest outgroups, the have dry fruits [19]. thecae remain separate at anthesis. However, it is equally In terms of the number of genera, this is the largest parsimonious to hypothesize that confluent anther the- tribe within subfamily Ajugoideae. Previous molecular cae are a synapomorphy of the clade comprising Aju- phylogenetic studies concentrated mainly on two genera, geae, Clerodendreae, and Teucrieae, with a subsequent Clerodendrum [34–36, 124] and Oxera [118, 119]. As a reversal at the base of Clerodendreae. result of the disintegration of the traditionally defined The traditionally delimited genus Caryopteris [5, 54, Clerodendrum, some genera (i.e., Volkameria, Ovieda, 127] is polyphyletic [9, 128] and species previously in- Rotheca) were resurrected [34–36, 123, 124]. Species re- cluded in Caryopteris have been distributed in six gen- lationships within those genera, however, remain uncer- era: Caryopteris, Discretitheca, Pseudocaryopteris, tain. In addition, relationships within the clade including Rubiteucris, Schnabelia, and Tripora, of which three Ovieda, Aegiphila, Clerodendrum, Tetraclea, Amasonia, were placed in tribe Ajugeae, two belong to tribe Teu- Kalaharia, and Volkameria, require further study. crieae, and one belongs to tribe Rotheceae. A sister- group relationship between Tripora and Pseudocaryop- teris was inferred in previous studies [54, 129, 130], but Subfamily Peronematoideae Bo Li, R.G. Olmstead & P.D. support values varied in different studies. The sister rela- Cantino tionship between the North American genus Trichos- Peronematoideae were recently established to accommo- tema and the East Asian genus Caryopteris was also date a well-supported clade comprising four small, reported in many studies [15, 16, 35, 36, 130, 131]. Al- mostly tropical Asian genera, Garrettia (1 sp.), Hymeno- though Ajuga is the largest genus in this tribe, no phylo- pyramis (7 spp.), Peronema (1 sp.), and Petraeovitex (8 genetic study has been carried out for the genus to date, spp.), which are sister to a larger clade formed by sub- and infrageneric relationships within this genus still need families Scutellarioideae, Cymarioideae, and Lamioideae further investigation. [19]. These four genera were previously placed in the subfamily Caryopteridoideae of Verbenaceae [5, 132, Tribe Clerodendreae Briq. 133] and were all transferred to Lamiaceae by Cantino Clerodendreae consist of ca. 350 species in ten genera: et al. [11], with Hymenopyramis placed in Viticoideae, Clerodendrum (ca. 150 spp.), Volkameria (30 spp.), Peronema and Petraeovitex in Teucrioideae, and Garret- Kalaharia Baill. (1 sp.), Amasonia L.f. (8 spp.), Tetraclea tia in Ajugoideae. However, all the four genera were A. Gray (2 spp.), Aegiphila Jacq. (120 spp.), Ovieda L. treated as incertae sedis in Harley et al.’s classification of (21 spp.), Oxera Labill. (21 spp.), Hosea Ridl. (1 sp.), and Lamiaceae [1]. probably Monochilus Fisch. & C.A. Mey. (2 spp.). Mono- In recent molecular phylogenetic studies, Garrettia chilus has not been included in any published molecular was first inferred to be sister to a clade comprising analysis, but based on a cladistic analysis of morpho- Scutellarioideae, Acrymia, Cymaria, and Lamioideae logical data, Cantino [9] suggested a close relationship [53], while the same sister relationship to an equiva- between Monochilus and Amasonia. Both genera usually lent clade of the Scutellarioideae-Cymaria-Lamioideae have alternate to subopposite leaves, a rare feature in clade (Acrymia was not sampled) was later found for Lamiaceae. Monochilus was not included in the molecu- a small well-supported clade comprised of Hymeno- lar results presented here but the presence of alternate pyramis, Petraeovitex,andPeronema [31, 44], as con- to subopposite leaves suggests that Monochilus should firmed in our phylogenomic trees (Fig. 1; Additional be treated within tribe Clerodendreae. However, this re- files 4, 5, 6, 7:Figs.S2,S3,S4,S5).WhenGarrettia, lationship needs to be tested using molecular evidence. Hymenopyramis, Petraeovitex,andPeronema were in- Clerodendreae are pan-tropical/subtropical in distribu- cluded in the same analysis, they grouped together in tion, predominantly distributed in the Americas, Africa, a highly supported clade that is sister to the Asia, and Pacific Oceania. A probable synapomorphy for Scutellarioideae-Cymarioideae-Lamioideae clade [19]. the tribe is a drupaceous fruit with four one-seeded pyr- Morphologically, the four genera are very heteroge- enes. In some species, the fruits split into four fleshy neous but do share some common traits as noted by schizocarps. A similar fruit type is found in Rotheca Chen et al. [44] and Li et al. [19]. Zhao et al. BMC Biology (2021) 19:2 Page 14 of 27

Subfamily Scutellarioideae (Dumort.) Caruel separate subfamily or as a distinct tribe within Lamioi- Scutellarioideae consist of ca. 390 species in five genera: deae; both options are acceptable based on the principle Holmskioldia Retz. (1 sp.), Wenchengia (1 sp.), Renschia of monophyly. However, as suggested by Bendiksby et al. Vatke (1 sp.), Tinnea Kotschy ex Hook. f. (19 sp.), and [53] and Chen et al. [44], the inclusion of Acrymia-Cym- Scutellaria L. (ca. 360 spp.) [1, 9, 121]. Species numbers aria within Lamioideae would make the subfamily mor- and distribution of these genera are extremely uneven. phologically heterogeneous and difficult to diagnose. Scutellaria is the largest and most widely distributed The apomorphy of axial monochasial cymes which de- genus, having a cosmopolitan distribution [1, 134, 135]. fines Cymarioideae is especially distinct and is not found Tinnea is much smaller and is distributed in tropical within Lamioideae. Thus, we concur with the approach and southern Africa. The monotypic genera Renschia, of Li et al. [19] and recognize Cymarioideae as a subfam- Wenchengia,andHolmskioldia are endemic to Somalia, ily here. Southeast Asia (Hainan Island of China, Vietnam), and subtropical Himalayan regions, respectively. Scutellarioi- Subfamily Lamioideae Harley deae is diagnosed by the following synapomorphic char- Lamioideae are the second largest subfamily within acters: pericarps with tuberculate or elongate processes Lamiaceae, containing about 1260 species in 62 genera, [136], high densities of xylem fibers in the calyces [137], with a near-cosmopolitan distribution, though concen- and thyrses with single-flowered cymes that form trated in Eurasia and northern to tropical Africa [52, 53, raceme-like inflorescences (but most species of Tinnea 69]. and Holmskioldia have cymose inflorescences). Considerable progress has been made in our under- Scutellarioideae had been thought to be sister to standing of subfamily Lamioideae in recent years. Since Lamioideae [31, 44], but with the separation of Cymar- Harley et al. [1], one genus has been established (Rydin- ioideae from the Lamioideae [19], Scutellarioideae is sis- gia Scheen & V.A. Albert [141]), four genera have been ter to the Cymarioideae-Lamioideae clade. Based on resurrected (Acanthoprasium (Benth.) Spenn. [53]; Beto- previous studies and our phylogenomic results, Tinnea nica [52]; Phlomoides Moench [142]; Pseudodictamnus and Holmskioldia are successive sister groups to Scutel- Fabr. [33]), eight genera have been reduced to synonyms laria, with Wenchengia sister to the rest of Scutellarioi- (Alajja Ikonn. and Sulaimania Hedge & Rech. f. [53]; deae [15, 16, 19, 31, 44, 66, 136]. However, relationships Lamiophlomis Kudô, Notochaete Benth., and Pseudere- within Scutellarioideae remain unresolved because mostachys Popov [142]; Eremostachys Bunge [28]; Bos- Renschia has never been included in a molecular phylo- trychanthera Benth. [43]; Stachyopsis Popov & Vved. genetic study. To date, four phylogenetic studies have fo- [143]), and Holocheila, which was formerly treated as cused on Scutellaria [66, 138–140], but none included a incertae sedis [1], has been shown to belong in Lamioi- comprehensive taxon sampling of the genus or of Scutel- deae [44]. Molecular phylogenies have also established larioideae as a whole. Thus, relationships within Scutel- that subfamily Cymarioideae is sister to Lamioideae [19]. laria still need to be addressed in future studies. A tribal classification of Lamioideae was the result of a molecular phylogeny based on cpDNA [52, 53]. The ten Subfamily Cymarioideae Bo Li, R.G. Olmstead & P.D. tribes have been corroborated as monophyletic groups Cantino using nuclear [143] and low-copy nuclear markers [67]. Cymarioideae were recently established to include two Four genera remained unplaced in the tribal classifica- small genera that have previously been considered incer- tion because they formed monogeneric clades [53, 67]; tae sedis [1], Acrymia (1 sp.) and Cymaria (2 spp.), however, two new tribes, i.e., Colquhounieae and Betoni- which are endemic to Southeast Asia. ceae, are proposed here to accommodate the genera Col- Bendiksby et al. [53] found that Acrymia and Cymaria quhounia and Betonica, respectively. The monotypic were the closest relatives of Lamioideae, which was sup- Roylea has still not been attributed to a tribe. Roylea ported by a subsequent study [44] but only with moder- groups within tribe Marrubieae in some nuclear-based ate support. Li et al. [19] further confirmed this phylogenies, but not in all and not in phylogenies based relationship with high support values and consequently on cpDNA data [33, 53, 67, 143]. To date, only two gen- established a new subfamily, Cymarioideae, to accommo- era, Metastachydium and Paralamium, have still not date the systematic position of the Acrymia-Cymaria been included in molecular phylogenetic studies of clade. In the present study, Cymaria dichotoma Benth. is Lamioideae, and their relationship with the other genera sister to Lamioideae in all analyses (Fig. 1; Additional remains enigmatic. files 4, 5, 6, 7: Figs. S2, S3, S4, S5). Regarding the systematic placement of the Acrymia- Tribe Pogostemoneae Briq. Cymaria clade, two different treatments are feasible [19]. Pogostemoneae consist of 11 genera as currently cir- The Acrymia-Cymaria clade could be treated as a cumscribed [44, 52, 53], including Achyrospermum Zhao et al. BMC Biology (2021) 19:2 Page 15 of 27

Blume (25 spp.), Anisomeles R. Br. (26 spp. [144]), Cra- suggested that small and relatively glossy nutlets, peri- niotome Rchb. (1 sp.), Colebrookea Sm. (1 sp.), carps (typically) lacking a sclerenchyma region [153, Comanthosphace S. Moore (4 spp.), Eurysolen Prain (1 154], generally long-exserted stamens with (usually) sp.), Holocheila (1 sp.), Leucosceptrum Sm. (1 sp.), bearded filaments, a (generally) weakly 2-lipped corolla, Microtoena Prain (19 spp.) [145], Pogostemon Desf. (80 and (generally) broad bracts are potentially useful mor- spp.) [146, 147], and Rostrinucula Kudô. (2 sp.), and all phological characters in defining the tribe. Further com- genera are monophyletic [44, 52, 53, 145, 148]. Most parative morphological studies combined with well- genera of the tribe are distributed in East Asia to South- supported phylogenetic trees based on extensive sam- east Asia, with three genera having a disjunct distribu- pling and additional nuclear loci will be necessary to de- tion between Asia and tropical Africa (Pogostemon, termine synapomorphies for this tribe. Achyrospermum, and Anisomeles). In addition, the monotypic genus Paralamium Dunn. is probably a Tribe Gomphostemmateae Scheen & Lindqvist member of Pogostemoneae based on the presence of Gomphostemmateae were established by Scheen et al. small glossy nutlets [53]. [52] to include three genera, Gomphostemma Wall. ex Pogostemoneae were established by Briquet [5] and Benth. (ca. 36 spp.), Chelonopsis Miq. (ca. 16 spp.), and originally included seven genera (Elsholtzia, Comantho- Bostrychanthera (2 spp.), that are distributed in temper- sphace, Keiskea, Pogostemon, Dysophylla Blume, Tetra- ate to tropical East Asia [1, 155]. Since then, the genus denia Benth., and Colebrookea). Later, Kudô [149] and Bostrychanthera was subsumed within Chelonopsis by Press [150] circumscribed Pogostemoneae in a broad Xiang et al. [43] based on morphological [156, 157] and sense, adding 11 genera to the tribe [11, 52, 150]. A molecular data [43] (see also Bongcheewin et al. [158]), number of taxonomic and molecular phylogenetic stud- thus leading to only two genera (Gomphostemma and ies [11, 19, 52, 53, 148, 150, 151] have indicated that six Chelonopsis) currently retained in this tribe. genera should be excluded from this tribe and that Dyso- Gomphostemmateae were shown to be sister to a large phylla should be merged with Pogostemon, as suggested group of Lamioideae in previous studies [52, 53], but by Hasskarl [152] and Press [150], leaving the present these results were equivocal due to suboptimal support total of 11 genera. values. Here, we find the same relationship but with Cantino [10] and Cantino et al. [11] proposed a sub- higher support values (Fig. 1). Possible synapomorphies family named Pogostemonoideae to include Colebrookea, for the tribe include pollen with branched columellae Comanthosphace, Leucosceptrum, Pogostemon, Rostrinu- [159] and fibers in the mesocarp [160, 161]. However, cula, Anisomeles,andEurysolen, but with hesitation re- pericarp structure has only been reported in a few spe- garding the two latter genera. Recent molecular cies, and it is unclear whether unexamined species share phylogenetic studies have shown that Pogostemonoideae these characters. Thus, future detailed morphological are sister to Lamioideae and have been included in that studies are needed. subfamily [52, 53]. Our results recover tribe Pogostemo- neae as sister to the clade contain all other members of Tribe Colquhounieae Lamioideae (Fig. 1). Previous studies based on plastid Colquhounieae are newly established here to accommo- DNA regions [52, 53] identified two well-supported date the enigmatic Colquhounia. The genus comprises clades within Pogostemoneae. One clade includes Eury- approximately five species endemic to the Himalayan solen, Leucosceptrum, Rostrinucula, Comanthosphace, massif from Nepal and north India to southwest China and Achyrospermum and is characterized by having dull and Vietnam. Morphologically, the genus is character- and glandular nutlets, and the sclerenchyma region in ized by having nutlets winged at the apex, which is rare the pericarp obsolete, indistinct, or absent. The second within subfamily Lamioideae [52]; besides Colquhounia, clade is composed of Colebrookea, Craniotome, Micro- only some species of Chelonopsis have this character. toena, Anisomeles, and Pogostemon. Within this clade, Based on trichome morphology, Hu et al. [162] classi- two subclades were recognized [53]. Colebrookea is the fied the genus into two sections, Colquhounia sect. Sim- only genus within the first subclade. This subclade is dis- plicipili C.Y. Wu & H.W. Li (including C. seguinii tinctive by possessing nutlets that are hairy and with Vaniot) and C. sect. Colquhounia (all remaining species), eglandular hairs at the apex, while the remaining genera but this classification was not supported by molecular formed a second subclade united by having glossy and phylogenetic results [43]. Although Scheen et al. [52] glabrous nutlets. Morphological studies focusing on and Bendiksby et al. [53] found that Colquhounia occu- traditionally defined Pogostemoneae (i.e., Pogostemonoi- pied a phylogenetically distinct position within Lamioi- deae; [153, 154]) identified some useful taxonomic char- deae, they kept the genus unclassified at the tribal level, acters. Subsequently, Scheen et al. [52], while not in part because only two species (C. coccinea Wall. and identifying any morphological synapomorphies, C. elegans Wall. ex Benth.) and only three markers Zhao et al. BMC Biology (2021) 19:2 Page 16 of 27

(trnL-trnF, rps16, and matK) were used for phylogenetic currently circumscribed. Recently, Giuliani and Bini reconstruction. In this study, five taxa were included for [174] found that Betonica possesses only peltate tri- analyses and they form a distinct clade within Lamioi- chomes, while Stachys has different types of large capi- deae. Thus, we describe this clade as a new tribe (see tate hairs that are lacking in Betonica. In addition, “Taxonomic treatment” below). Giuliani and Bini [174] also found that peltate trichomes of Betonica species have unusual secretions composed of Tribe Synandreae Raf. flavonoids and essential oils and suggested that Betonica Synandreae were recircumscribed by Scheen et al. [163] should be considered a genus of its own. to include the following five genera: Brazoria Englm. & An early molecular phylogenetic analysis of Stachys s.l. A. Gray (3 spp.), Macbridea Elliott ex Nutt. (2 spp.), [39] based on both plastid and nuclear DNA sequence Physostegia Benth. (12 spp.), Synandra Nutt. (1 sp.), and data demonstrated that the type species of Betonica, B. Warnockia M.W. Turner (1 sp.). The only morpho- officinalis L. (as Stachys officinalis (L.) Trevis.), fell out- logical synapomorphy for the tribe is a raceme-like in- side of the clade that contained the remainder of Stachys florescence with sessile or very shortly pedicellate including the type species, S. sylvatica L. Perhaps not be- flowers [52, 163]. All five genera are characterized by ing aware of this, Harley et al. [1] retained Betonica in having villous stamen filaments, but this is also found in synonymy under Stachys. Since then, further molecular some members of tribe Pogostemoneae (e.g., Pogoste- phylogenetic evidence has corroborated the distinctness mon, Anisomeles) and tribe Stachydeae [1](Chamae- of Betonica [52, 53, 69]. Based on results from a compre- sphacos Schrenk ex Fisch. & C.A. Mey). hensive plastid phylogeny of Lamioideae that included Previous studies involving cpDNA, nrDNA, and low- five species of Betonica, Scheen et al. [52] suggested copy nuclear markers failed to adequately discern the Betonica should be resurrected from synonymy under position of Synandreae within Lamioideae [52, 53, 69, Stachys. The five species of Betonica formed a strongly 163, 164]. Our results provide strong support for the supported clade sister to Galeopsis, the sister relation- placement of Synandreae, with the caveat that only one ship, however, receiving low statistical support. This representative was included (Fig. 1). Only two lamioid phylogenetic result was corroborated by a follow-up tribes include species with a North American distribu- study with more taxa and additional genetic markers tion, Synandreae and Stachydeae. The two tribes are not [53]. The monophyly and distinctness of Betonica has closely related [52, 53, 69] and therefore represent separ- received support also from nuclear data [69], with a ate dispersals into North America [163]. weakly supported sister relationship to tribe Synandreae. Since Betonica so far has remained unclassified at the Tribe Betoniceae tribal level, and the genus seems to lack a clear affinity Betoniceae are newly established here to accommodate to any other lamioid taxon, we propose herein that the the phenetically and genetically isolated genus Betonica Betonica clade be recognized at the tribal level (see in Lamioideae. There are nine currently accepted Beto- “Taxonomic treatment” below). nica species, three of which include 2–6 subspecific taxa The intrageneric classification of Betonica into three distributed throughout Europe reaching Central Asia sections [175, 176] (i.e., B. sect. Betonica, B. sect. Folio- and Northwest Africa [165]. Betonica has repeatedly sae (Krestovsk. & Lazkov) Lazkov, and B. sect. Macrosta- been included in, and excluded from, the genus Stachys chya (R. Bhattacharjee) Krestovsk.), has not received L. Some authors have treated Betonica as a distinct statistical support by any so far published molecular genus [166–171], while Bentham [90] and Briquet [5] phylogenies [e.g., 52, 53]. Betonica alopecuros L., how- placed Betonica in its own section within Stachys. In the ever, receives support as sister to the remaining species most recent morphological classification of Stachys, in these studies. The distinctness of B. alopecuros is also Bhattacharjee [172] recognized Betonica as a subgenus supported by morphology: yellow corollas with bifid (S. subg. Betonica (L.) Bhattacharjee) within Stachys, de- upper lip and annulate corolla tubes [172, 177]. A more fined by prominent sterile rosettes, usually unbranched comprehensive study of this genus is needed. flowering shoots arising from an axillary bud of the root- stock, and deeply crenate to serrate leaf margins; fea- Tribe Galeopseae (Dumort.) Vis. tures that Betonica shares with the Stachys sections In the present study, we propose resurrection of the Eriostomum (Hoffmanns. & Link) Dumort. and Ambleia tribe Galeopseae to accommodate the phenetically and Benth. Yet, Bhattacharjee [172] mentions that S. subg. genetically isolated genus Galeopsis in Lamioideae. Betonica is divergent in the nature of the calyx (sessile Dumortier [167] established subtribe Galeopsinae (as calyx) and bracteoles (with a broad hardened base). “Galeopsideae”) within the Stachydeae and included the Tomas-Bárberán et al. [173] points to differences in two genera, Galeopsis and Lamium L. Later, Visiani phytochemistry between Betonica and Stachys species, as [178] elevated subtribe Galeopsinae to the rank of tribe Zhao et al. BMC Biology (2021) 19:2 Page 17 of 27

(as “Galeopsideae” [98]) but included only Galeopsis. represents the most likely maternal parent to G. tetrahit Galeopsis represents a morphologically highly distinct (G. pubescens being the paternal parent). Galeopsis subg. genus within subfamily Lamioideae, characterized by Ladanum comprises the following five species: G. erect annual herbs with two conical protuberances near ladanum L., G. nana Otsch., G. pyrenaica Bartl., G. reu- the base of the anterior lip of the corolla and anthers de- teri Rchb. f., G. segetum Neck. Species within G. subg. hiscing by two valves, of which the upper is fimbriate [1, Ladanum have proven indistinguishable in DNA phylo- 179]. genetic analyses involving nuclear (NRPA2, 5S-NTS) Galeopsis comprises 10 currently accepted species, two and chloroplast (matK, psbA-trnH, rps16, trnL-trnF, and subspecies, and six documented hybrids [165]. The trnS-trnG) DNA regions (M. Bendiksby, unpubl.). Mor- genus is distributed in temperate Eurasia with a center phologically, however, they appear highly distinct, and of species richness in Europe [180]. Strong support for AFLP data (genomic fingerprint) group accessions ac- the monophyly of Galeopsis was obtained in two recent cording to species [M. Bendiksby, unpubl.]. Hence, the molecular phylogenetic studies of Lamioideae that in- species of G. subg. Ladanum have probably diverged re- cluded three [52] and eight [53] species of Galeopsis, cently and the multilocus data suffers from incomplete respectively. lineage sorting. Phylogenetically, Galeopsis holds a rather isolated, yet uncertain, position and has remained unclassified at the tribal level [52, 53, 69]. In both Scheen et al. [52] and Tribe Stachydeae Dumort. Bendiksby et al. [53], Galeopsis was weakly supported as Cosmopolitan Stachydeae are the largest and taxonomic- sister to Betonica, which in turn was sister to tribe Sta- ally most challenging alliance of all recognized tribes in chydeae, with even weaker support. Hence, a close rela- subfamily Lamioideae [29, 30, 52, 53, 69]. Stachydeae tionship to Lamium and Lamiastrum Heist. ex Fabr., have previously been the subject of several molecular with which Galeopsis had been classified in most trad- phylogenetic investigations [30, 37–39, 52, 53, 185]. itional classifications (e.g., subtribe Galeopsidinae Lindqvist and Albert [39] revealed that three genera en- Dumort.) based on the shared feature of a swollen cor- demic to Hawaiian (dry fruited Haplostachys (A. Gray) olla tube, was discarded. The sister relationship between W.F. Hillebr., fleshy fruited Phyllostegia Benth., and Ste- Betonica and Galeopsis received some support from nogyne Benth.) as well as the genera Prasium L., Phlomi- other sources of data. The two genera share the same doschema (Benth.) Vved., and Sideritis L. are nested base chromosome number [181](x = 8), and flavonoid within the large genus Stachys. Both Prasium with fleshy p-coumaroyl glucosides are present in both Betonica and schizocarp and Phlomidoschema, which is characterized G. subg. Galeopsis [182]. The placement of Galeopsis in by a small corolla and branched hairs, are monotypic the nuclear PPR phylogeny by Roy and Lindqvist [69], [1]. In contrast, Stachys comprises about 275 species and however, does not support a sister relationship to Beto- Sideritis comprises about 125 species [1]. This paraphyly nica, and Galeopsis falls out largely unresolved in their of Stachys was corroborated by Scheen et al. [52] who Lamioideae phylogeny. Although our current plastome- showed that the Asian genera Chamaesphacos, Suzukia based phylogeny corroborates a sister relationship be- Kudô, and Thuspeinanta T. Durand also are embedded tween Galeopsis and Betonica, support values remain within Stachys and that the monotypic genus Melittis L. low (Fig. 1). Based upon this phylogenetic uncertainty, represents the sister to all other Stachydeae. In a later the lack of support from nuclear data, and a goal of work, Bendiksby et al. [53] added Hypogomphia Bunge achieving taxonomic stability, we resurrect the tribe to the list of taxa nested within Stachys. Morphologic- Galeopseae to encompass the single genus Galeopsis. ally, the annuals Chamaesphacos, Hypogomphia, and Reichenbach [169] divided Galeopsis into two subgen- Thuspeinanta are characterized by 1–3-flowered cymes era, G. subg. Galeopsis and G. subg. Ladanum Rchb. and narrow nutlets, while Suzukia is recognized by a Subgenus Galeopsis, is readily distinguished from creeping habit and racemose inflorescences [1]. Thus, 12 G. subg. Ladanum by the presence of rigid hairs and genera and ca. 470 species are currently recognized in swollen stem nodes in the former. The division of the Stachydeae, but generic realignments are needed to re- genera into two equally sized subgenera is supported by flect phylogenetic relationships. Scheen et al. [52] found phytochemistry [182], crossing experiments [183], and no non-molecular synapomorphies for this diverse tribe, molecular phylogenetics [53, 184]. Galeopsis subg. but listed the following characteristics as common Galeopsis comprises the following five species: G. bifida among its members: calyx campanulate or weakly 2- Boenn., G. pubescens Besser., G. speciosa Mill., G. tetra- lipped, calyx lobes often spiny, calyx throat often hairy, hit L., and G. sulphurea Jord. According to molecular corolla strongly 2-lipped, anterior pair of stamens bend- analyses by Bendiksby et al. [184], the latter appears to ing outwards after pollination, and nutlets usually api- represent a valid species, distinct from G. speciosa, and cally rounded. Zhao et al. BMC Biology (2021) 19:2 Page 18 of 27

Besides some studies focusing on certain groups, such China and Japan), with some species of Paraphlomis ex- as Hawaiian [39, 185] and New World Stachys [30, 69], a tending to Southeast Asia [1, 193]. comprehensive phylogenetic study of Stachydeae based on multiple loci analyses was performed by Salmaki Tribe Phlomideae Mathiesen et al. [29]. Analyses of nuclear ribosomal (nrITS) and Based on the most recent molecular phylogenetic study plastid DNA data corroborated the monophyly of the of Phlomideae [28], the tribe now consist of only two tribe, with Melittis as sister to all remaining Stachydeae. genera: Phlomis L. (ca. 50–90 spp.) and Phlomoides (ca. Salmaki et al. [29] suggested the phylogenetic name 150–170 spp.). “Eurystachys Y. Salmaki & M. Bendiksby” for the clade Phlomideae were established by Mathiesen in Scheen including all genera attributed to Stachydeae except et al. [52], in which six genera were recognized in the Melittis. Although the plastid DNA markers provided tribe: Eremostachys, Lamiophlomis, Notochaete, Phlomis, well-supported backbone resolution in the Eurystachys Phlomoides, and Pseuderemostachys. Phlomideae are clade, the nrITS phylogenetic tree recovered several usually characterized by having calyx lobes abruptly nar- groups with relatively poorly supported and short rowed to a narrow apex and expanded at the corolla branches [29]. Therefore, detailed conclusions on the margins that are bearded and densely pubescent outside phylogenetic relationships in the Eurystachys clade and have branched hairs [52]. Mathiesen et al. [142] later needed using additional nuclear markers. reduced Pseuderemostachys, Lamiophlomis, and one spe- Recently, phylogenetic relationships in the Eurystachys cies of Notochaete (N. hamosa Benth.) to synonyms of clade utilizing two additional nuclear ribosomal DNA se- Phlomoides. Combining multilocus molecular phylogen- quences (nrETS and 5S-NTS) provided high resolution etic analyses and morphological evidence, Salmaki et al. allowing recognition of 12 well-supported clades within [28] continued to show that Eremostachys, Notochaete, the Eurystachys clade, which also were recovered in the and Paraeremostachys Adylov, Kamelin & Makhm previous phylogenetic analyses using plastid DNA se- should all be transferred to Phlomoides. Thus, the num- quences [186]. The 12 clades were formally named in ber of recognized genera in Phlomideae was reduced to the Eurystachys clade following a PhyloCode nomencla- two, i.e., Phlomis and Phlomoides. Species of Phlomis are ture [187] and provided the basis for a future rank-based shrubs or subshrubs with simple leaves, laterally com- classification of Stachydeae with two options: (1) split- pressed, flattened, sickle-shaped, but not fringed or in- ting the Eurystachys clade into 12 individual genera, cised upper corolla lips, and with nutlet pericarps each based on a pre-existing genus name and redefined possessing a sclerenchyma region (indistinct in a few to encompass additional taxa, but without clear morpho- species). In contrast, Phlomoides are herbaceous with logical apomorphies; or (2) lumping of all these formal simple or laciniate to pinnatisect leaves and with upper clades into a broadly defined Stachys, including widely corolla lips that are arch-shaped, and always hairy or recognized and morphologically well-defined segregates fringed-incised, but not laterally compressed or flat- such as Prasium and Sideritis [186]. Clearly, more stud- tened, and have pericarps lacking a sclerenchyma region ies using various sources of evidence are needed to clar- [194]. Phlomis have a mostly circum-Mediterranean dis- ify the taxonomic borders in this tribe. A micro- tribution, while the centers of diversification of Phlo- morphological approach [153, 159–161, 188–192]ata moides include Central Asia, the Iranian highlands, and global scale may provide a promising supplement to the China [28, 142, 195]. more traditionally applied macro-morphological approaches. Tribe Leonureae Dumort. Tribe Paraphlomideae Bendiksby Leonureae were recircumscribed by Scheen et al. [52] Paraphlomideae were established by Bendiksby et al. and Bendiksby et al. [53] based on phylogenetic and [53] to accommodate Matsumurella Makino (5 spp.), morphologic data. They are comprised of 80 species in Ajugoides Makino (1 sp.), and Paraphlomis (Prain) Prain six genera: Chaiturus Willd. (1 sp.), Lagochilus Bunge ex (ca. 25 spp.), together which have been found to form a Benth. (45 spp.), Leonurus L. (24 spp.), Panzerina Soják distinct lineage within Lamioideae. Though the tribe has (2 spp.), Loxocalyx Hemsl. (3 spp.), and Lagopsis (Bunge no clear synapomorphy, it can be distinguished from ex Benth.) Bunge (5 spp.). The tribe is distributed pri- other tribes of Lamioideae by the following set of char- marily in Central Asia. Phylogenetic studies have shown acters: herbs or subshrubs, indumentum of simple hairs, that Lagopsis and Leonurus are poly- or paraphyletic actinomorphic calyx, corolla (1/3) with hairy upper lip [53]. Possible morphological synapomorphies for the but scarcely bearded along the margin, included sta- tribe are short stamens included in the corolla tube and mens, and an apically truncate ovary [1, 53, 193]. Most more or less palmate venation and lobing of the leaves. species of the tribe are restricted to East Asia (south The genus Loxocalyx lacks these characters but shares Zhao et al. BMC Biology (2021) 19:2 Page 19 of 27

zygomorphic calyces with longer abaxial lobes with subcontinent to Queensland, Australia [199]. With a few many Leonureae. exceptions, members of Leucadeae have a calyx that is distinctly zygomorphic with secondary lobes and a Tribe Marrubieae Vis. bearded margin of the upper lip of the corolla [52]. The Marrubieae, with about 91 species, consist mostly of latter character is also found in the genus Phlomoides non-aromatic herbs or subshrubs, with thyrsoid inflores- [52]. The monophyly of Leucadeae has been corrobo- cences, few- to many-flowered cymes, widely campanu- rated using low-copy nuclear data [69], although only a late to rotate calyces often with secondary calyx lobes, small but representative selection of species was in- zygomorphic and 2-lipped corollas, and included or cluded in this study. shortly exserted stamens [1]. The tribe is distributed One molecular phylogeny has included a wide repre- from Europe to west and central Asia as well as North sentation of species from all six genera, but only cpDNA and South Africa with the highest number of species in markers were analyzed [199]. The large genus Leucas, southern Europe and North Africa [33]. with more than 100 species occurring on dry or dis- The taxonomy and generic delimitations within Mar- turbed ground in tropical to southern Africa and tropical rubieae have been controversial [33, 52, 53, 196–198]. and subtropical parts of Asia [1], was shown to be para- Marrubieae contained three genera, Ballota L., Marru- phyletic with respect to Acrotome and Leonotis, Isoleu- bium L., and Moluccella L. based on Scheen et al. [52]. cas, and Otostegia [199]. Only a few of the Asian species Later, Bendiksby et al. [53] showed that the two species of Leucas were included, but they formed a clade separ- of B. sect. Acanthoprasium Benth. (B. integrifolia Benth., ate from the remaining Leucas [199]. More data are B. frutescens (L.) Woods) form a clade separate from the needed, including low-copy nuclear markers, before remaining species of Ballota. Therefore, Bendiksby et al. taxonomic changes can be proposed. [53] resurrected the genus Acanthoprasium as proposed The genus Otostegia, as traditionally circumscribed, (but not formalized) by Scheen et al. [52]. The mono- was clearly polyphyletic [199]. To make Otostegia mono- typic Sulaimania Hedge & Rech. f. was recovered as a phyletic, the genus Rydingia was described to accommo- member of the Moluccella clade and reduced to syn- date four Asian species, one species was transferred to onymy of Moluccella [53]. In a recent phylogenetic study Isoleucas, and one species was transferred to Moluccella of tribe Marrubieae using four plastid and one nuclear [141]. Since then, an additional four species of Otostegia DNA locus (ITS), B. sect. Beringeria (Neck.) Benth. was have also been transferred to Moluccella [53] (see also raised to generic rank, as Pseudodictamnus Fabr. [33]. the discussion on tribe Marrubieae). Thus, the recircum- Therefore, the tribe now comprises five genera: Acantho- scribed Otostegia is reduced to ca. eight species, most of prasium (2 spp.), Ballota (3 spp.), Marrubium (ca. 50) which are endemic to Africa [53], with O. fruticosa spp., Moluccella (8 spp.), and Pseudodictamnus (28 spp.) (Forssk.) Schweinf. ex Penz. extending to the Arabian [33]. Peninsula [200]. Members of the genus Acanthoprasium are shrubby Molecular phylogenies have resolved Rydingia as sister and woody, have long spiny bracteoles, and occur in to the rest of Leucadeae, with this relationship recovered Europe, while species of Pseudodictamnus are herb- based on cpDNA [52, 53] and low-copy nuclear DNA aceous, have leafy bracteoles, and are predominantly [69]. However, more data are still needed to resolve the Mediterranean-African in distribution [33, 91]. Ballota generic boundaries of the paraphyletic genus Leucas in as now circumscribed includes herbaceous species cov- relation to Acrotome, Isoleucas, Leonotis,andOtostegia. ered by simple trichomes and are distributed from Eur- ope to West Asia (including also the Mediterranean) [33]. Marrubium was also recircumscribed recently to Tribe Lamieae Coss. & Germ. include B. deserti (de Noé) Jury, Rejdali & A.J.K. Grif- Lamieae are comprised of four genera: Lamium (includ- fiths. There are around 50 species assigned to this genus, ing Wiedemannia Fisch. & C.A. Mey and Lamiastrum; which are characterized by a bifid upper corolla lip and ca. 25 spp.), Eriophyton Benth. (including Alajja; ca. 8 distributed from Macaronesia to temperate Eurasia. spp.), Stachyopsis (4 spp.), and possibly Menitskia (Kres- tovsk.) Krestovsk. (1 sp.). These genera are widely dis- Tribe Leucadeae Scheen & Ryding tributed in the temperate and subtropical regions of Leucadeae were established by Scheen et al. [52] and in- Europe, Asia, and Northern Africa. Five East Asian spe- clude ca. 134 species in six genera: Acrotome Benth. ex cies of Galeobdolon and Lamium chinense Benth. were Endl. (8 spp.), Isoleucas O. Schwartz (2 spp.), Leonotis transferred into the genus Matsumurella in tribe Para- (Pers.) R. Br. (9 spp.), Leucas R. Br. (ca. 100 spp.), Otoste- phlomideae by Bendiksby et al. [53]. Possible morpho- gia Benth. (ca. 8 spp.), and Rydingia (4 spp.). These gen- logical synapomorphies for the tribe are hairy anthers era are distributed from Africa through the Indian (except for Lamium galeobdolon L., L. flexuosum Ten., Zhao et al. BMC Biology (2021) 19:2 Page 20 of 27

L. orvala L. and some other species in the genus Erio- Colquhounieae consist of one genus and approxi- phyton) and nutlets subtruncate or truncate at apex. mately five species, occurring from Nepal, across north Ryding [201]includedWiedemannia within La- India to southwest and central China and Vietnam. mium, and Harley et al. [1]includedLamiastrum (syn. Galeobdolon)inLamium. Scheen et al. [52]de- Rotheceae fined Lamieae to consist of a single genus Lamium C.L. Xiang, Bo Li & R.G. Olmstead, trib. nov. Type: (including Lamiastrum and Wiedemannia). Scheen Rotheca Raf. et al. [52] and Bendiksby et al. [202] both found a Shrubs, subshrubs, and perennial herbs. Leaves simple, clade comprised of Lamium, Lamiastrum,andWiede- opposite, or whorled with 3–4 leaves per node, often mannia, but did not have sufficient sampling to as- toothed. Flowers often in terminal and/or axillary cymes. sess monophyly of Lamium, e.g., if Lamiastrum and Calyx actinomorphic, 5-lobed or truncate. Corolla ± Wiedemannia were excluded. Subsequent studies, with zygomorphic, expanding abruptly on lower side only; 5 more complete sampling of Lamium,foundLamias- lobes ± unequal, anterior corolla lobe frequently much trum to be nested within Lamium [203, 204]. larger than the other four, limb in bud asymmetrical. Bendiksby et al. [53] also determined that two other Stamens 4, didynamous to subequal, long-exserted; an- genera, Eriophyton (including Alajja and three species thers usually basifixed (occasionally approaching versa- of Lamium)andStachyopsis,shouldbeincludedin tile). Ovary unlobed in flower but becoming imperfectly tribe Lamieae. Bendiksby et al. [143]foundthatSta- 4-lobed during fruit development. Style terminal, stigma chys tibetica Vatke (= Menitskia tibetica (Vatke) Kres- lobes frequently unequal. Drupes (2–) 4-lobed, meso- tovsk.) did not belong in Stachys (tribe Stachydeae), carp ± fleshy, endocarp separated into 4 stones or 2 pairs but was most closely related to Stachyopsis in of stones. Lamieae. Morphologically, however, S. tibetica has an The tribe contain four genera, Rotheca (60 spp.), Glos- intermediate position between Stachyopsis and Erio- socarya (13 spp.), Discretitheca (1 sp.), and Karomia (9 phyton.TheyexpandedEriophyton to include Sta- spp.), and are distributed in tropical southern Asia to chyopsis and S. tibetica,inordertomakeEriophyton southern Africa, and Australia (Queensland). monophyletic. Lazkov and Sennikov [176]statedthat the genus Stachyopsis is similar to Eriophyton but dif- Betoniceae fers in the habit, shape of leaves (oblong-ovate vs. Bendiksby & Salmaki, trib. nov. Type: Betonica L. broadly rhomboid-ovate), and shorter flower tube Perennial herbs. Leaves deeply crenate-dentate. which is enclosed within the calyx; therefore, they Flowering stems unbranched, lateral to rootstock, verti- suggested that the genus Stachyopsis should retain its cillasters condensed (rarely remote), 16–20-flowered. generic status. At the same time, they resurrected Bracteoles scarious or herbaceous, apex spinescent, base Menitskia to accommodate S. tibetica as Menitskia broad and hardened. Flowers sessile, median lobe of tibetica.ThegenusMenitskia differs from Eriophyton lower corolla lip emarginate. Calyx sessile, ± regular. An- and Stachyopsis by its narrower posterior corolla lip, ther cells subparallel to parallel. stiffer bracteoles, and often deeply crenate to lobed Betoniceae are monotypic comprising the genus Beto- leaves [143, 205]. nica with about 10 species distributed in western Eurasia. Taxonomic treatment Colquhounieae Conclusions C.L. Xiang, Bo Li & R.G. Olmstead, trib. nov. Type: Col- This is the first study to use plastome data to estimate quhounia Wall. family-wide relationships within Lamiaceae. We demon- Shrubs erect or ascending. Stems and branches ter- strate that increased taxon sampling in concert with ete, with simple and/or branched hairs. Leaves phylogenomic analyses based on plastome sequence data toothed, petiolate; inflorescence thyrsoid, pedunculate provides superior support and resolution at both deep to subsessile; cymes 1–5-flowered. Calyx tubular- and shallow nodes relative to previous studies and offers campanulate, 10-veined, 5-lobed, lobes often equal. new insights into phylogenetic relationships among and Corolla strongly 2-lipped, 4-lobed (1/3), often purple, between tribes and subfamilies of Lamiaceae. The mono- sometimes spotted; posterior lip moderately long, phyly of all 12 subfamilies is corroborated, and we hooded with upcurved margins, anterior lip slightly recognize a total of 22 tribes within Lamiaceae, three of subequally 3-lobed, corolla tube strongly dilated dis- which are newly established here (i.e. Colquhounieae, tally; stamens 4, not exserted from corolla, thecae ± Rotheceae, and Betoniceae). This study provides a de- confluent; stigma lobes unequally 2-cleft. Nutlets nar- tailed summary of the taxonomic history, generic and rowly obovoid-oblong, winged at apex. species diversity, morphology, synapomorphies, and Zhao et al. BMC Biology (2021) 19:2 Page 21 of 27

distribution for each tribe and subfamily, representing and the start and stop codons were manually adjusted the most comprehensive overview of Lamiaceae since by comparison with the plastome of Salvia miltiorrhiza Harley et al. [1]. The classification presented herein is Bunge [213] (HF586694). The online tRNAscan-SE web the most definitive tribal-level taxonomy of the mint servers [214] were used to confirm the tRNA genes. Cir- family to date, and the robust phylogenetic backbone of cular plastome maps were drawn using the Organellar- Lamiaceae reconstructed here provides an extendable GenomeDRAW tool [215]. For 19 species, the plastomes dataset for future studies on mint family classification, were assembled from RNA-seq data and only contigs biogeography, character evolution, and diversification. were obtained. Bowtie2 [216] was then used to map contigs to the reference sequences extracted from S. mil- Materials and methods tiorrhiza [213]. Taxon sampling In this study, plastomes of 50 taxa were newly sequenced Sequence alignment and dataset generation and 61 taxa were reassembled from the sequence read Since noncoding regions can be variable even among archive (SRA) database; others were acquired from pre- species and are often difficult to align across a family as vious studies [66, 67, 206, 207] or downloaded from large as Lamiaceae, only 79 protein-coding genes were NCBI (https://www.ncbi.nlm.nih.gov; Additional file 8: used for phylogenetic analyses. Alignments of individual Table S3). In total, the ingroup sampling included 170 loci were performed using the MAFFT v.7.308 [217] plu- taxa (175 accessions), 79 genera, and represented all 15 gin in Geneious v.11.0.3 [212] with G-INS-I algorithm, currently recognized tribes and all 12 subfamilies within and the final alignments were manually adjusted in Lamiaceae [19, 51]. Twenty-two species from five fam- PhyDE v.0.9971 [218]. ilies of Lamiales (Mazaceae, Orobanchaceae, Phryma- Since the plastome is uniparentally inherited in most ceae, Paulowniaceae, and Wightiaceae) were selected as angiosperms and generally does not undergo recombin- outgroups based on phylogenetic results of previous ation, sequences of the 79 coding genes were studies [18, 22, 208]. Voucher specimens of the newly concatenated in our study to generate a supermatrix of sequenced taxa (Table 1) were deposited at the Herbar- all coding regions (CR). Removal of problematic aligned ium of Kunming Institute of Botany, Chinese Academy regions may result in a better resolved phylogeny [219]; of Sciences (KUN). therefore, ambiguously aligned positions (e.g., characters of uncertain homology among taxa and single-taxon in- DNA isolation and sequencing sertions; see [31, 46]) were removed manually in our DNA was extracted from healthy and fresh leaves frozen analyses to construct the “Coding region manual” dataset in liquid nitrogen or dried in silica gel using the CTAB (CRM, Additional file 3: Table S2). protocol of Doyle and Doyle [209] and sheared into ca. Additional matrices for the 79 genes were constructed 300 bp fragments using a Covaris M220 Focused- based on (1) the 1st and 2nd codon positions (CR12); (2) ultrasonicator. Libraries for paired-end (PE) Illumina se- only the 3rd codon positions (CR3); and (3) the degener- quencing were constructed from fragmented genomic ated coded sequences (dePCS) generated using Degen DNA following the standard protocol of manufacture v.1.4 (http://www.phylotools.com/). Thus, a total of five (NEBNext® Ultra II™DNA Library Prep Kit for Illumina®) datasets (CR, CRM, CR12, CR3, dePCS) were used in and sequenced from both ends of 150 bp fragments on subsequent analyses. the Illumina HiSeq 2000 platform (Illumina, San Diego, CA, USA) at BGI Genomics (BGO-Shenzhen, China). Phylogenetic analyses Approximately 2–10 GB of raw data was generated with Phylogenetic trees based on all datasets were built by 150 bp paired-end read lengths. two approaches including Bayesian inference (BI) ana- lysis and maximum likelihood (ML) analysis. jModelTest Plastome assembly and annotation v.2.1.4 [220] was used to determine the best-fit models Quality control of raw sequence reads was carried out for nucleotide sequences for BI analyses. using FastQC toolkit [210](http://www.bioinformatics. Bayesian analyses were executed using MrBayes v.3.2.2 babraham.ac.uk/projects/fastqc) with the parameter set [221]. Four iterations of 50,000,000 generations were run as Q ≥ 25 to acquire high-quality clean reads. The de on four chains, sampling every 1000 generations on the novo assembling of the plastome was implemented in Cyberinfrastructure for Phylogenetic Research Science the GetOrganelle pipeline [68], in which plastome reads (CIPRES) Gateway v.3.3 server [222](http://www.phylo. were extracted from total genomic reads and then org/). Default priors, unlinked parameter estimates, and SPAdes v.3.10 [211] was used for assembly. For those best-fit models suggested by jModelTest v.2.1.4 [220] for plastomes we can acquire complete sequences, genome each dataset were used for each iteration. Convergence annotation was performed using Geneious v.11.0.3 [212], of runs was accepted when the average standard Zhao et al. BMC Biology (2021) 19:2 Page 22 of 27

deviation of split frequencies (ASDSF) dropped below sequence (dataset dePCS). A, phylogram showing branch lengths, where 0.01. Tracer v.1.6.0 [223] was used to inspect the conver- tip names are absent follow the same order as shown in B. Scale bar gence of model parameters and check whether the represents the mean number of nucleotide substitutions per site. B, ≥ maximum likelihood bootstrap support values are shown above the values of effective sample size (ESS) were 200. A branches. majority-rule consensus tree was created from the runs, Additional file 8: Table S3. List of taxa sampled with information after a 25% burn-in. All resulting trees with nodal sup- related to taxonomy, GenBank accession numbers, references, and port values were visualized and edited in FigTree v.1.4.2 vouchers. Herbarium acronyms follow Index Herbariorum [227]. [224]. ML analyses were performed using RAxML v.8.2.9 Acknowledgements The authors thank Dr. Miao Sun, Dr. Qing-Qing Zhan, Dr. Zhuo Zhou, Dr. [225] as implemented in the XSEDE interface of CIPRES Ren-Bin Zhu, Paul Carmen (Australian National Botanic Gardens), and Dr. Jian [222]. The GTRCAT model was used for analyses and Huang for their assistance in sample collection. bootstrapping; bootstrap iterations (–#|–N) were set to 1000, and other parameters used the CIPRES default Authors’ contributions CLX, BL, FZ, YPC, HP, and RGO conceived this research. FZ, YPC, BTD, TCW, settings. FC, BL, and CLX collected materials. FZ performed the experiments. FZ, YPC, We defined branches with posterior probabilities (PP) and XCL analyzed the data. All the authors wrote the manuscript. All authors < 0.90 and bootstrap values (BS) < 70% as weakly sup- read and approved the final version of manuscript. ported, PP = 0.90–0.95 and BS = 70%–80% as moderately Funding supported, and PP ≥ 0.95 and BS ≥ 80% as strongly sup- This study was in part supported by Yunnan Fundamental Research Projects ported [107]. The alignments and ML tree are deposited (Grant No. 2019FI009) and “Ten Thousand Talents Program of Yunnan” (Grant “ ” at TreeBase with study #S26639 (http://treebase.org/ No. YNWR-QNBJ-2018-279) awarded to CLX, the CAS Light of West China program to CLX and YPC, the Australian Biological Resources Study National treebase-web/phylows/study/TB2:S26639?x-access- Taxonomy Research Grant Program (Grant No. RG19-17) awarded to TCW, code=bb02a4c5bc226f4604690ea0f21ccd41&format= CLX, and BL, NSF DEB-1655611 awarded to BTD, and the Postdoctoral html)[226]. Research Program (TUBITAK-BIDEB 2219) awarded to FC. Availability of data and materials All the newly sequenced and annotated plastomes in the present study Supplementary Information were submitted to the National Center for Biotechnology Information (NCBI) The online version contains supplementary material available at https://doi. database with accession numbers MT473738–MT473786 (Table 1). Other org/10.1186/s12915-020-00931-z. plastomes analyzed were acquired from previous studies [66, 67, 206, 207]or downloaded from NCBI (https://www.ncbi.nlm.nih.gov; Additional file 1: Figure S1. Gene maps of the complete chloroplast Additional file 8; Table S3). The alignments and ML tree are deposited at genomes newly sequenced in this study. Genes inside and outside of the TreeBase with study #S26639 (http://treebase.org/treebase-web/phylows/ circle are transcribed in the clockwise and counterclockwise directions, study/TB2:S26639?x-access-code=bb02a4c5bc226f4604690ea0f21ccd41 respectively. Genes belonging to different functional categories are color- &format=html)[226]. coded. Additional file 2: Table S1. Features of newly sequenced plastomes. Ethics approval and consent to participate Not applicable. Additional file 3: Table S2. Excluded ambiguous sites for 79 genes of coding regions (dataset CRM). Consent for publication Additional file 4: Figure S2. Phylograms inferred from ML analysis of Not applicable. concatenated nucleotide sequences of 79 protein-coding genes (dataset CR). A, phylogram showing branch lengths, where tips names are absent Competing interests follow the same order as shown in B. Scale bar represents the mean number of nucleotide substitutions per site. B, maximum likelihood boot- The authors declare that they have no competing interests. strap support values and Bayesian inference posterior probabilities are shown above and below the branches, respectively. Author details 1CAS Key Laboratory for Plant Diversity and Biogeography of East Asia, Additional file 5: Figure S3. Phylograms inferred from ML analysis of Kunming Institute of Botany, Chinese Academy of Sciences, Kunming rd concatenated nucleotide sequences of the 3 codon positions (dataset 650201, China. 2Center of Excellence in Phylogeny of Living Organisms, CR3). A, phylogram showing branch lengths, where tip names are absent Department of Plant Science, College of Science, University of Tehran, P.O. follow the same order as shown in B. Scale bar represents the mean Box 14155-6455, Tehran, Iran. 3Department of Biology, University of Nebraska number of nucleotide substitutions per site. B, maximum likelihood at Kearney, Kearney, NE 68849, USA. 4National Herbarium of New South bootstrap support values and Bayesian inference posterior probabilities Wales, Australian Institute of Botanical Science, Royal Botanic Gardens & are shown above and below the branches, respectively. Domain Trust, Sydney, Australia. 5Museum of Archaeology, University of Additional file 6: Figure S4. Phylograms inferred from ML analysis of Stavanger, NO-4036 Stavanger, Norway. 6Department of Biology, Faculty of concatenated nucleotide sequences of the 1st and 2nd codon positions Arts and Sciences, Kırıkkale University, Kırıkkale, Turkey. 7Department of (dataset CR12). A, phylogram showing branch lengths, where tips Botany, University of Wisconsin–Madison, Madison, WI 53706, USA. names are absent follow the same order as shown in B. Scale bar 8Department of Botany, Natural History Museum Vienna, Burgring 7, 1010 represents the mean number of nucleotide substitutions per site. B, Wien, Austria. 9NTNU University Museum, Norwegian University of Science maximum likelihood bootstrap support values and Bayesian inference and Technology, 7491 Trondheim, Norway. 10Natural History Museum, posterior probabilities are shown above and below the branches, University of Oslo, Oslo, Norway. 11State Key Laboratory of Systematic & respectively. Evolutionary Botany, Institute of Botany, Chinense Academy of Sciences, Xiangshan, Beijing 100093, China. 12Research Centre of Ecological Sciences, Additional file 7: Figure S5. Phylograms inferred from ML analysis of concatenated nucleotide sequences of the degeneracy nucleotide College of Agronomy, Jiangxi Agricultural University, Nanchang 330045, China. 13Department of Biology, University of Washington, Seattle, WA, USA. Zhao et al. BMC Biology (2021) 19:2 Page 23 of 27

Received: 21 July 2020 Accepted: 19 November 2020 26. Pastore JFB, Harley RM, Forest F, Paton AJ, Van den Berg C. Phylogeny of the subtribe Hyptidinae (Lamiaceae tribe Ocimeae) as inferred from nuclear and plastid DNA. Taxon. 2011;62:1217–329. 27. Drew BT, Sytsma KJ. Phylogenetics, biogeography, and staminal evolution in References the tribe Mentheae (Lamiaceae). Am J Bot. 2012;99:933–53. 1. Harley RM, Atkins S, Budantsev AL, Cantino PD, Conn BJ, Grayer R, et al. 28. Salmaki Y, Zarre S, Ryding O, Lindqvist C, Scheunert A, Bräuchler C, et al. Labiatae. In: Kadereit JW, editor. The families and genera of vascular plants, Phylogeny of the tribe Phlomideae (Lamioideae: Lamiaceae) with special – vol. 7. Berlin: Springer Verlag; 2004. p. 167 275. focus on Eremostachys and Phlomoides: new insights from nuclear and 2. Olmstead RG. A synoptical classification of the Lamiales, version 2.6.2. 2016. chloroplast sequences. Taxon. 2012;65:161–79. http://depts.washington.edu/phylo/Classification.pdf. Accessed 25 Feb 2020. 29. Salmaki Y, Zarre S, Ryding O, Lindqvist C, Bräuchler C, Heubl G, et al. (last updated 12 April, 2016). Molecular phylogeny of tribe Stachydeae (Lamiaceae subfamily Lamioideae). 3. APG IV. An update of the Angiosperm Phylogeny Group classification for Mol Phylogenet Evol. 2013;69:535–51. the orders and families of flowering plants: APG IV. Bot J Linn Soc. 2016;181: 30. Roy T, Chang T-H, Lan T, Lindqvist C. Phylogeny and biogeography of New 1–20. World Stachydeae (Lamiaceae) with emphasis on the origin and 4. Bentham G. Verbenaceae and Labiatae. In: Bentham G, Hooker JD, editors. diversification of Hawaiian and South American taxa. Mol Phylogenet Evol. Genera plantarum. London: Reeve; 1876. p. 1131–223. 2013;69:218–38. 5. Briquet J. Verbenaceae, Labiatae. In: HGA E, KAE P, editors. Die Natürlichen 31. Chen Y-P, Drew BT, Li B, Soltis DE, Soltis PS, Xiang C-L. Resolving the Pflanzenfamilien. Berlin: Engelmann, W; 1897. p. 132–375. phylogenetic position of Ombrocharis (Lamiaceae), with reference to the 6. Erdtman G. Pollen morphology and plant taxonomy. IV. Labiatae, molecular phylogeny of tribe Elsholtzieae. Taxon. 2016;65:123–36. Verbenaceae and Avicenniaceae. Svensk Bot Tidskr. 1945;39:277–85. 32. Paton AJ, Mwanyambo M, Culham A. Phylogenetic study of Plectranthus, 7. Wunderlich R. Ein Vorschlag zu einer natürlichen Gliederung der Labiaten Coleus and allies (Lamiaceae): taxonomy, distribution and medicinal use. Bot auf Grund der Pollenkörner, der Samenentwicklung und des reifen Samens. J Linn Soc. 2018;188:355–76. Oester Bot Zeit. 1967;114:383–483. 33. Siadati S, Salmaki Y, Mehrvarz SS, Heubl G, Weigned M. Untangling the 8. Cantino PD, Sanders RW. Subfamilial classification of Labiatae. Syst Bot. generic boundaries in tribe Marrubieae (Lamiaceae: Lamioideae) using 1986;11:163–85. nuclear and plastid DNA sequences. Taxon. 2018;67:770–83. 9. Cantino PD. Evidence for a polyphyletic origin of the Labiatae. Ann Mo Bot 34. Steane DA, Scotland RW, Mabberley DJ, Wagstaff SJ, Reeves PA, Olmstead Gard. 1992;79:361–79. RG. Phylogenetic relationships of Clerodendrum s.l. (Lamiaceae) inferred from 10. Cantino PD. Toward a phylogenetic classification of the Labiatae. In: Harley chloroplast DNA. Syst Bot. 1997;22:229–43. RM, Reynolds T, editors. Advances in Labiatae Science. London: Royal Botanic Gardens, Kew; 1992. p. 27–37. 35. Steane DA, Scotland RW, Mabberley DJ, Olmstead RG. Molecular systematics 11. Cantino PD, Harley RM, Wagstaff SJ. Genera of Labiatae: status and of Clerodendrum (Lamiaceae): ITS sequences and total evidence. Am J Bot. – classification. In: Harley R, Reynolds T, editors. Advances in Labiatae science. 1999;86:98 107. London: Royal Botanic Gardens, Kew; 1992. p. 511–22. 36. Steane DA, De Kok RPJ, Olmstead RG. Phylogenetic relationships between 12. Junell S. Zur Gynäzeummorphologie und Systematik der Verbenaceen und Clerodendrum (Lamiaceae) and other Ajugoid genera inferred from nuclear – Labiaten, nebst Bemerungen über ihre Samenentwicklung. Symb Bot Upsal. and chloroplast DNA sequences data. Mol Phylogenet Evol. 2004;32:39 45. 1934;4:1–129. 37. Barber JC, Francisco-Ortega J, Santos-Guerra A, Turner KG, Jansen RK. Origin 13. Kadereit JW. Lamiales: introduction and conspectus. In: Kadereit JW, editor. of Macaronesian Sideritis L. (Lamioideae: Lamiaceae) inferred from nuclear – The Families and Genera of Vascular Plants, Vol. 7. Berlin: Springer Verlag; and chloroplast sequence datasets. Mol Phylogent Evol. 2002;23:293 306. 2004. p. 1–8. 38. Barber JC, Finch CC, Francisco-Ortega J, Santos-Guerra A, Jansen RK. 14. Wink M, Kaufmann M. Phylogenetic relationships between some members Hybridization in Macaronesian Sideritis (Lamiaceae): Evidence from of the subfamily Lamioideae (Family Labiatae) inferred from nucleotide incongruence of multiple independent nuclear and chloroplast sequence – suquences of the rbcL gene. Bot Acta. 1996;109:139–48. datasets. Taxon. 2007;56:74 88. 15. Wagstaff SJ, Olmstead RG. Phylogeny of Labiatae and Verbenaceae inferred 39. Lindqvist C, Albert VA. Origin of the Hawaiian endemic mints within the – from rbcL sequences. Syst Bot. 1997;22:165–79. North American Stachys (Lamiaceae). Am J Bot. 2002;89:1709 24. 16. Wagstaff SJ, Hickerson L, Spangler R, Reeves PA, Olmstead RG. Phylogeny in 40. Walker JB, Sytsma KJ, Treutlein J, Wink M. Salvia (Lamiaceae) is not Labiatae s.l., inferred from cpDNA sequences. Plant Syst Evol. 1998;209:265– monophyletic: implications for the systematics, radiation, and ecological 74. specializations of Salvia and tribe Mentheae. Am J Bot. 2004;91:1115–25. 17. Schäferhoff B, Fleischmann A, Fischer E, Albach DC, Borsch T, Heubl G, et al. 41. Walker JB, Drew BT, Sytsma KJ. Unravelling species relationships and Towards resolving Lamiales relationships: insights from rapidly evolving diversification within the iconic California Floristic Province sages (Salvia chloroplast sequences. BMC Evol Biol. 2010;10:352. subgenus Audibertia, Lamiaceae). Syst Bot. 2015;40:826–44. 18. Refulio-Rodriguez NF, Olmstead RG. Phylogeny of Lamiidae. Am J Bot. 2014; 42. Bramley GLC, Forest F, De Kok RPJ. Troublesome tropical mints: Re- 101:287–99. examining generic limits of Vitex and relations (Lamiaceae) in South East 19. Li B, Cantino PD, Olmstead RG, Bramley GL, Xiang C-L, Ma Z-H, et al. A Asia. Taxon. 2009;58:500–10. large-scale chloroplast phylogeny of the Lamiaceae sheds new light on its 43. Xiang C-L, Zhang Q, Scheen A-C, Cantino PD, Funamoto T, Peng H. Molecular subfamilial classification. Sci Rep. 2016;6:34343. phylogenetics of Chelonopsis (Lamiaceae: Gomphostemmateae) as inferred 20. Mint Evolutionary Genomics Consortium. Phylogenomic mining of the from nuclear and plastid DNA and morphology. Taxon. 2013;62:375–86. mints reveals multiple mechanisms contributing to the evolution of 44. Chen Y-P, Li B, Olmstead RG, Cantino PD, Xiang C-L, Liu E-D. Phylogenetic chemical diversity in Lamiaceae. Mol Plant. 2018;11:1084–96. placement of the enigmatic genus Holocheila (Lamiaceae) inferred from 21. Soltis DE, Smith SA, Cellinese N, Wurdack KJ, Tank DC, Brockington SF, et al. plastid DNA sequences. Taxon. 2014;63:355–66. Angiosperm phylogeny: 17 genes, 640 taxa. Am J Bot. 2011;98:704–30. 45. Drew BT, Cacho NI, Sytsma KJ. The transfer of two rare monotypic genera, 22. Liu B, Tan Y-H, Liu S, Olmstead R, Ming D-Z, Chen Z-D, et al. Phylogenetic Neoeplingia and Chaunostoma,toLepechinia (Lamiaceae), and notes on relationships of Cyrtandromoea and Wightia revisited: a new tribe in their conservation. Taxon. 2014;63:831–42. Phrymaceae and a new family in Lamiales. J Syst Evol. 2020;58:1–17. 46. Drew BT, González-Gallegos JG, Xiang C-L, Kriebel R, Drummond CP, Walker 23. Paton AJ, Springate D, Suddee S, Otieno D, Grayer RJ, Harley MM, et al. JB, et al. Salvia united: The greatest good for the greatest number. Taxon. Phylogeny and evolution of basils and allies (Ocimeae, Labiatae) based on 2017;66:133–45. three plastid DNA regions. Mol Phylogenet Evol. 2004;31:277–99. 47. González-Gallegos JG, Drew B, Cuevas-Guzmán R. Novelties on the 24. Conn BJ, Streiber N, Brown EA, Heywood MJ, Olmstead RG. Infrageneric distribution of Lepechinia flammea (Lamiaceae), rediscovery of L. glomerata phylogeny of Chloantheae (Lamiaceae) based on chloroplast ndhF and in Jalisco, Mexico, and their phylogenetic position within Lepechinia. Bot Sci. nuclear ITS sequence data. Aust Syst Bot. 2009;22:243–56. 2015;93:1–14. 25. Bräuchler C, Meimberg H, Heubl G. Molecular phylogeny of Menthinae 48. Salmaki Y, Kattari S, Heubl G, Bräuchler C. Phylogeny of non-monophyletic (Lamiaceae, Nepetoideae, Mentheae) – taxonomy, biogeography and Teucrium (Lamiaceae: Ajugoideae): implications for character evolution and conflicts. Mol Phylogenet Evol. 2010;55:501–23. taxonomy. Taxon. 2016;65:802–22. Zhao et al. BMC Biology (2021) 19:2 Page 24 of 27

49. Will M, Claßen-Bockhoff R. Time to split Salvia s.l. (Lamiaceae)–New insights 74. Munir AA. A taxonomic revision of Chloanthaceae trib. Physopsideae from Old World Salvia phylogeny. Mol Phylogenet Evol. 2017;109:33–58. Brunonia. 1978;1:407–692. 50. Hu G-X, Takano A, Drew BT, Liu E-D, Soltis DE, Soltis PS, et al. Phylogeny and 75. Brown R. Prodromus Florae Novae Hollandiae. London: J. Johnson & Co; staminal evolution of Salvia (Lamiaceae, Nepetoideae) in East Asia. Ann Bot. 1810. 2018;122:649–68. 76. Reichenbach HGL. Conspectus regni vegetabilis per gradus naturales 51. Li B, Olmstead RG. Two new subfamilies in Lamiaceae. Phytotaxa. 2017;313: evoluti. Leipzig: Carolum Cnobloch; 1828. 222–6. 77. Von Mueller FJH. Nova genera et species aliquot rariores in plagis Australiae. 52. Scheen A-C, Bendiksby M, Ryding O, Mathiesen C, Albert VA, Lindqvist C. Hooker’s J Bot Kew Gard Misc. 1857;9:14–24. Molecular phylogenetics, character evolution and suprageneric classification 78. Hutchinson J. The Families of Flowering Plants. 2nd ed. London: Oxford of Lamioideae (Lamiaceae). Ann Mo Bot Gard. 2010;97:191–219. University Press; 1959. 53. Bendiksby M, Thorbek L, Scheen A-C, Lindqvist C, Ryding O. An updated 79. Munir AA. A taxonomic revision of the genus Hemiphora (Chloanthaceae). J phylogeny and classification of Lamiaceae subfamily Lamioideae. Taxon. Adel Bo Gard. 1978;1:161–6. 2011;60:471–84. 80. Eichler H. Supplement to J.M. Black’s Flora of South Australia. 2nd edition. 54. Xiang C-L, Zhao F, Cantino PD, Drew BT, Li B, Liu E-D, et al. Molecular Adelaide: Govt. Printer; 1965. systematics of Caryopteris (Lamiaceae) and its allies with reference to the 81. Munir AA. A taxonomic revision of the genus Chloanthes (Chloanthaceae). J molecular phylogeny of subfamily Ajugoideae. Taxon. 2018;67:376–94. Adel Bo Gard. 1977;1:83–106. 55. Moore MJ, Bell CD, Soltis PS, Soltis DE. Using plastid genome-scale data to 82. Munir AA. A taxonomic revision of the genus Cyanostegia (Chloanthaceae). resolve enigmatic relationships among basal angiosperms. Proc Natl Acad Brunonia. 1978;1:45–67. Sci U S A. 2007;104:19363–8. 83. Munir AA. A taxonomic revision of the genus Pityrodia (Chloanthaceae). J 56. Olmstead RG, Bedoya AM. Whole genomes: the holy grail. A commentary Adel Bo Gard. 1979;2:1–138. on: ‘molecular phylogenomics of the tribe Shoreeae (Dipterocarpaceae) 84. Blackall WE, Grieve BJ. How to know Western Australian wildflowers: a key using whole plastidgenomes’. Ann Bot. 2019;123:iv–v. to the flora of the temperate regions of Western Australia. Nedlands: 57. Lin C-P, Huang J-P, Wu C-S, Hsu C-Y, Chaw S-M. Comparative chloroplast University of Western Australia Press; 1965. genomics reveals the evolution of Pinaceae genera and subfamilies. 85. Cronquist A. The evolution and classification of flowering plants. Aylesbury: Genome Biol Evol. 2010;2:504–17. Hazel Watson & Viney; 1968. 58. Moore MJ, Soltis PS, Bell CD, Burleigh JG, Soltis DE. Phylogenetic analysis of 86. Wilson TC, Radunz EA, Chen S, Conn BJ, Henwood MJ. A new genus and 83 plastid genes further resolves the early diversification of eudicots. Proc species for Chloantheae (Lamiaceae). Aust Syst Bot. 2020; In press. Natl Acad Sci U S A. 2010;107:4623–8. 87. Guerin GR. Evidence for polyphyly in Hemigenia and Microcorys (Lamiaceae: 59. Xi Z, Ruhfel BR, Schaefer H, Amorim AM, Sugumaran M, Wurdack KJ, et al. Westringieae). Aust Syst Bot. 2008;21:313–25. Phylogenomics and a posteriori data partitioning resolve the Cretaceous 88. Huck RB. Overview of pollination biology in the Lamiaceae. In: Harley RM, angiosperm radiation Malpighiales. Proc Natl Acad Sci U S A. 2012;109:17519–24. Reynolds T, editors. Advances in Labiatae Science. London: Royal Botanic 60. Stull GW, De Stefano RD, Soltis DE, Soltis PS. Resolving basal lamiid Gardens, Kew; 1992. p. 167–81. phylogeny and the circumscription of Icacinaceae with a plastome-scale 89. Wilson TC, Conn BJ, Henwood MJ. Great expectations: correlations between data set. Am J Bot. 2015;102:1794–813. pollinator assemblages and floral characters in Lamiaceae. Int J Plant Sci. 61. Zhang S-D, Jin J-J, Chen S-Y, Chase MW, Soltis DE, Li H-T, et al. 2017;178:170–87. Diversification of Rosaceae since the late Cretaceous based on plastid 90. Wilson TC, Conn BJ, Henwood MJ. Molecular phylogeny and systematics of phylogenomics. New Phytol. 2017;214:1355–67. Prostanthera (Lamiaceae). Aust Syst Bot. 2012;25:341–52. 62. Zhang R, Wang Y-H, Jin J-J, Stull GW, Bruneau A, Cardoso D, et al. 91. Bentham G. Labiatarum genera et species. London: J. Ridgeway & Sons; 1834. Exploration of plastid phylogenomic conflict yields new insights into the 92. Guerin GR. Nutlet morphology in Hemigenia R.Br. and Microcorys R.Br. deep relationships of Leguminosae. Syst Biol. 2020;69:613–22. (Lamiaceae). Plant Syst Evol. 2005;254:49–68. 63. Yu X-Q, Gao L-M, Soltis DE, Soltis PS, Yang J-B, Fang L, et al. Insights into the 93. Hedge IC. Labiatae. In: Ali SI, Nasir YJ, editors. Flora of Pakistan, vol. 192. historical assembly of East Asian subtropical evergreen broadleaved forests revealed Karachi: University of Karachi; 1990. p. 310. by the temporal history of the tea family. New Phytol. 2017;215:1235–48. 94. Kaufmann M, Wink M. Molecular systematics of the Nepetoideae 64. Yang X-Y, Wang Z-F, Luo W-C, Guo X-Y, Zhang C-H, Liu J-Q, et al. Plastomes (Lamiaceae): Phylogenetic implications from rbcL gene sequences. Z of Betulaceae and phylogenetic implication. J Syst Evol. 2019;57:508–18. Naturforsch C. 1994;49:635–45. 65. Xiang C-L, Dong H-J, Landrein S, Zhao F, Yu W-B, Soltis DE, et al. Revisiting 95. Wagstaff SJ, Olmstead RG, Cantino PD. Parsimony analysis of cpDNA restriction the phylogeny of Dipsacales: new insights from phylogenomic analyses of site variation in subfamily Nepetoideae (Labiatae). Am J Bot. 1995;82:886–92. complete plastomic sequences. J Syst Evol. 2020;58:103–17. 96. Martin AC. The comparative internal morphology of seeds. Ame Mid Natur. 66. Zhao F, Li B, Drew BT, Chen Y-P, Wang Q, Yu W-B, et al. Leveraging 1946;36:513–660. plastomes of comparative analysis and phylogenomic inference within 97. Ryding O. The distribution and evolution of myxocarpy in Lamiaceae. In: Scutellarioideae (Lamiaceae). Plos One. 2020;15:e0232602. Harley RM, Reynolds T, editors. Advances in Labiatae Science. London: Royal 67. Zhao F, Drew BT, Chen Y-P, Hu G-X, Li B, Xiang C-L. The chloroplast Botanic Gardens, Kew; 1992. p. 85–96. genome of Salvia species: genomic characterization and phylogenetic 98. Li P, Qi Z-C, Liu L-X, Ohi-Toma T, Lee J, Hsieh TH, et al. Molecular analysis. Int J Plant Sci. 2020;181:812–30. phylogenetics and biogeography of the mint tribe Elsholtzieae 68. Jin J-J, Yu W-B, Yang J-B, Song Y, de Pamphilis CW, Yi T-S, et al. (Nepetoideae, Lamiaceae), with an emphasis on its diversification in East GetOrganelle: a fast and versatile toolkit for accurate de novo assembly of Asia. Sci Rep. 2017;7:2057. organelle genomes. Genome Biol. 2020;21:241. 99. Sanders RW, Cantino PD. Nomenclature of the subdivisions of the 69. Roy T, Lindqvist C. New insights into evolutionary relationships within the Lamiaceae. Taxon. 1984;33:64–72. subfamily Lamioideae (Lamiaceae) based on pentatricopeptide repeat (PPR) 100. Funamoto T, Xiang C-L, Ogawa M, Peng H. A comparative study of nuclear DNA sequences. Am J Bot. 2015;102:1721–35. chromosome characters in four species of Elsholtzia Will. (Lamiaceae) in 70. Guignard ML. Recherches sur la développement de la graine et en Japan and China. Chromosome Bot. 2012;7:119–23. particulier du tégument seminal: Labiées. J Bot. 1893;7:241–50. 101. Mayta-Anco LF, Molinari-Novoa EA, Raju AJS. A new genus within the 71. Olmstead RG, Reeves PA, Lepschi BJ. Confirmation of a monophyletic Elsholtziae (Lamiaceae) from southeast Asia. Weberbauerella. 2016;1:1–7. Chloanthoideae (Lamiaceae) comprising tribes Chloantheae and 102. Paton AJ, Ryding O. Hanceola, Siphocranion and Isodon and their position in Prostanthereae. Lamiales Newsletter. 1998;6:7–10. the Ocimeae (Labiatae). Kew Bull. 1998;53:723–31. 72. Godden GT, Kinser TJ, Soltis PS, Soltis DE. Phylotranscriptomic analyses 103. Harley RM, Pastore JFB. A generic revision and new combinations in the reveal asymmetrical gene duplication dynamics and signatures of ancient Hyptidinae (Lamiaceae), based on molecular and morphological evidence. polyploidy in mints. Genome Biol Evol. 2019;11:3393–408. Phytotaxa. 2012;58:1–55. 73. Conn BJ, Henwood MJ, Streiber N. Synopsis of the tribe Chloantheae and 104. Paton AJ, Mwanyambot M, Govaerts RHA, Smitha K, Suddee S, Phillipson PB, new nomenclatural combinations in Pityrodia s.lat. (Lamiaceae). Aust Syst et al. Nomenclatural changes in Coleus and Plectranthus (Lamiaceae): a tale Bot. 2011;24:1–9. of more than two genera. PhytoKeys. 2019;129:1–158. Zhao et al. BMC Biology (2021) 19:2 Page 25 of 27

105. Harley RM, Paton AJ, Ryding O. New synonym and taxonomic changes in 133. Chen S-L, Gilbert MG. Verbenaceae. In: Wu Z-Y, Raven PH, editors. Flora of Laibatae. Kew Bull. 2003;58:485–9. China, Vol. 17. Beijing: Science Press; St. Louis: Missouri Botanical Garden 106. Zhong J-S, Li J, Li L, Conran JG, Li H-W. Phylogeny of Isodon (Schrad. ex Press; 1994. pp. 1–49. Benth.) Spach (Lamiaceae) and related genera inferred from nuclear 134. Paton AJ. A global taxonomic investigation of Scutellaria (Labiatae). Kew ribosomal ITS, trnL–trnF region, and rps16 intron sequences and Bull. 1990;45:399–450. morphology. Syst Bot. 2010;35:207–19. 135. Paton AJ. The phytogeography of Scutellaria L. Notes Roy Bot Gard 107. Chen Y-P, Zhu X-X, Zhao F, Feng H-Z, Paton A, Xiang C-L. Molecular and Edinburgh. 1990;46:345–9. morphological evidence for a new species of Siphocranion (Lamiaceae) 136. Li B, Xu W-X, Tu T-Y, Wang Z-S, Olmstead RG, Peng H, et al. Phylogenetic from the Sino-Vietnamese border. Phytotaxa. 2019;425:1–18. position of Wenchengia (Lamiaceae): a taxonomically enigmatic and 108. Walker JB, Sytsma KJ. Staminal evolution in the genus Salvia (Lamiaceae): critically endangered genus. Taxon. 2012;61:392–401. molecular phylogenetic evidence for multiple origins of the staminal lever. 137. Ryding O. Amount of calyx fibres in Lamiaceae, relation to calyx structure, Ann Bot. 2007;100:375–91. phylogeny and ecology. Plant Syst Evol. 2007;268:45–58. 109. Drew BT, Sytsma KJ. Testing the monophyly and placement of Lepechinia in 138. Chiang Y-C, Huang B-H, Liao P-C. Diversification, biogeographic pattern, and the tribe Mentheae (Lamiaceae). Syst Bot. 2011;36:1038–49. demographic history of Taiwanese Scutellaria species inferred from nuclear 110. Drew BT. Phylogenetics and biogeography of Lepechinia (Lamiaceae), and and chloroplast DNA. Plos One. 2012;7:e50844. evolutionary studies within the Mentheae Tribe. Ph.D. dissertation. Madison: 139. Zhao F, Liu E-D, Peng H, Xiang C-L. A new species of Scutellaria University of Wisconsin; 2011. (Scutellarioideae, Lamiaceae) from Sichuan Province in southwest China. 111. Rouy G. Compectus de la Flore de France. Paris: Rue de tournon; 1927. PeerJ. 2017;5:e3624. 112. Moon HK, Vinckier S, Smets E, Huysmans S. Palynological evolutionary 140. Safikhani K, Jamzad Z, Saeidi H. Phylogenetic relationships in Iranian trends within the tribe Mentheae with special emphasis on subtribe Scutellaria (Lamiaceae) based on nuclear ribosomal ITS and chloroplast trnL- Menthinae (Nepetoideae: Lamiaceae). Plant Syst Evol. 2008;275:93–108. F DNA data. Plant Syst Evol. 2018;304:1077–89. 113. Li J-C, Zhang J-W, Zhang D-G, Deng T, Volis S, Sun H, et al. Phylogenetic 141. Scheen A-C, Albert VA. Nomenclatural and taxonomic changes within the position of the Chinese endemic genus Heterolamium: a close relative of Leucas clade (Lamioideae; Lamiaceae). Syst Geogr Pl. 2007;77:229–38. subtribe Nepetinae (Lamiaceae). J Jap Bot. 2017;92:12–9. 142. Mathiesen C, Scheen A-C, Lindqvist C. Phylogeny and biogeography of the 114. Takhtajan A. Flowering Plants. Netherlands: Springer; 2009. lamioid genus Phlomis (Lamiaceae). Kew Bull. 2011;66:83–99. 115. Thorne RF. Classification and geography of the flowering plants. Bot Rev. 143. Bendiksby M, Salmaki Y, Bräuchler C, Ryding O. The generic position of 1992;58:225–348. Stachys tibetica Vatke and amalgamation of the genera Eriophyton and 116. Cronquist A. The evolution and classification of flowering plants. New York: Stachyopsis (Lamiaceae subfam. Lamioideae). Plant Syst Evol. 2014;300:961–71. Columbia University Press; 1988. 144. Bean AR. A taxonomic revision of Anisomeles R.Br. (Lamiaceae). 117. Reveal JL. An outline of a classification scheme for extant flowering plants. Austrobaileya. 2015;9:321–81. Phytoneuron. 2012;37:1–211. 145. Wang Q. A monograph of the genus Microtoena (Lamiaceae). Beijing: 118. Barrabé L, Karnadi-Abdelkader G, Ounémoa J, De Kok RPJ, Robert N, Gâteblé Science Press; 2018. G. Recircumscription of Oxera (Lamiaceae: Ajugoideae) to include Faradaya 146. Bhatti GR, Ingrouilie M. Systematics of Pogostemon (Labiatae). Bull Nat Hist based on molecular and anatomical data. Bot J Linn Soc. 2015;179:693–791. Mus Lond. (Bot.). 1997;27:77–147. 119. Barrabé L, Lavergne S, Karnadi-Abdelkader G, Drew BT, Birnbaum P, Gâteblé 147. Yao G, Deng Y-F, Ge X-J. A taxonomic revision of Pogostemon (Lamiaceae) G. Changing ecological opportunities facilitated the explosive diversification from China. Phytotaxa. 2015;200:1–67. of New Caledonian Oxera (Lamiaceae). Syst Biol. 2018;68:460–81. 148. Yao G, Drew BT, Yi T-S, Yan H-F, Yuan Y-M, Ge X-J. Phylogenetic 120. Thorne RF. The calssificaiton and geography of the flowering plants: relationships, character evolution and biogeographic diversification of Dicotyledons of the class Angiospermae. Bot Rev. 2000;66:441–647. Pogostemon s.l. (Lamiaceae). Mol Phylogenet Evol. 2016;98:184–200. 121. Abu-Asab MS, Cantino PD. Phylogenetic implications of pollen morphology 149. Kudô Y. Labiatarum sino-japonicarum prodromus. Mem Fac Taihoku Imp in Tribe Ajugeae (Labiatae). Syst Bot. 1993;18:100–22. Univ. 1929;2:1–332. 122. Dumortier BC. Analyse des Familles des Plantes. Tournay: J. Casterman; 1829. 150. Press JR. Taxonomic studies in the Labiatae tribe Pogostemoneae. Bull Nat 123. Steane DA, Mabberley DJ. Rotheca (Lamiaceae) revived. Novon. 1998;8:204–6. Hist Mus Lond (Bot). 1982;10:1–89. 124. Yuan Y-W, Mabberley DJ, Steane DA, Olmstead RG. Further disintegration 151. Keng H. Labiatae. In: CGGJ VS, editor. Flora Malesiana ser. 1, 8(3). Leyden: and redefinition of Clerodendrum (Lamiaceae): implications for the Noordhoff International Publishing; 1978. p. 301–94. understanding of the evolution of an intriguing breeding strategy. Taxon. 152. Hasskarl JC. Plantarum genera et species novae aut reformatae Javenses. 2010;59:125–33. Flora 25 Beibl. 1842;2(1–2):1–56. 125. Charles-Dominique T, Davies TJ, Hempson GP, Bezeng BZ, Daru BH, 153. Ryding O. Pericarp structure and phylogeny of Lamiaceae subfamily Kabongo RM, et al. Spiny plants, mammal browsers, and the origin of Pogostemonoideae. Nord J Bot. 1994;14:59–63. African savannas. Proc. Natl Acad Sci U S A. 2016;113:E5572–9. 154. Ryding O. Pericarp structure and phylogeny of the Lamiaceae-Verbenaceae- 126. Huang MJ, Crawford DJ, Freudenstein JV, Cantino PD. Systematics of complex. Pl Syst Evol. 1995;198:101–41. Trichostema (Lamiaceae): evidence from ITS, ndhF, and morphology. Syst 155. Xiang C-L, Liu E-D, Peng H. A key to the genus Chelonopsis (Lamiaceae) and Bot. 2008;33:437–46. two new combinations: C. rosea var. siccanea and C. souliei var. cashmerica 127. Chen S-L, Guo R-L. Caryopteris. In: Chen S-L, Pei J, editors. Flora Reipublicae comb. nov. Nord J Bot. 2008;26:31–4. Popularis Sinicae, vol. 65(1). Beijing: Science Press; 1982. p. 194–208. 156. Xiang C-L, Dong Z-H, Peng H, Liu Z-W. Trichome micromorphology of the 128. Rimpler H, Winterhalter C, Falk U. Cladistic analysis of the subfamily East Asiatic genus Chelonopsis (Lamiaceae) and its systematic implications. Caryopteridoideae Briq. and related taxa of Verbenaceae and Lamiaceae Flora. 2010;205:434–41. using morphological and chemical characters. In: Harley RM, Reynolds T, 157. Xiang C-L, Funamoto T, Evangelista EV, Zhang Q, Peng H. Pollen editors. Advances in Labiatae Sciences. London: Royal Botanic Gardens, Kew; morphology of the East Asiatic genus Chelonopsis (Lamioideae: Lamiaceae) 1992. p. 39–54. and allied genera, with reference to taxonomic implications and potential 129. Huang MJ. Systematics of Trichostema L. (Lamiaceae) and phylogenetic pollination ecology. Plant Biosyst. 2013;147:620–8. relationships with its disjunct taxa in Asia. Ph.D. Dissertation. The Ohio State 158. Bongcheewin B, Grayer RJ, Porter EA, Kite GC, Simmonds MSJ, Ingrouille MJ, University; 2002. et al. A chemossystematic study of the genus Gomphostemma and related 130. Shi S-H, Du Y-Q, Boufford DE, Gong X, Huang Y-L, He H-H, et al. Phylogenetic genera (Lamiaceae). Biochem Syst Ecol. 2014;57:305–16. position of Schnabelia, a genus endemic to China: Evidence from sequences of 159. Abu-Asab MS, Cantino PD. Systematic implications of pollen morphology in cpDNA matK gene and ITS regions. Chin Sci Bull. 2003;48:1576–80. subfamilies Lamioideae and Pogostemonoideae (Labiatae). Ann Mo Bot 131. Cantino PD, Wagstaff ST, Olmstead RG. Caryopteris (Lamiaceae) and the Gard. 1994;81:653–86. conflict between phylogenetic and pragmatic considerations in botanical 160. Ryding O. Pericarp structure in the subtribe Melittidinae (Lamiaceae- nomenclature. Syst Bot. 1998;23:369–86. Lamioideae) and its systematic implications. Bot Jahrb Syst. 1994;115:547–55. 132. P’ei C, Chen S-L. Verbenaceae. In: Chen SL, Pei J, editors. Flora Reipublicae 161. Ryding O. Pericarp structure in the tribe Prasieae (Lamiaceae-Lamioideae) Popularis Sinicae, Vol. 65(1). Beijing: Science Press; 1982. p. 1–208. and its systematic implications. Bot Jahrb Syst. 1994;116:391–9. Zhao et al. BMC Biology (2021) 19:2 Page 26 of 27

162. Hu G-X, Balangcod TD, Xiang C-L. Trichome micromorphology of the 190. Salmaki Y, Zarre S, Jamzad Z. Nutlet micromorphology and its systematic Chinese-Himalayan genus Colquhounia (Lamiaceae), with emphasis on implication in Stachys L. (Lamiaceae) in Iran. Feddes Repert. 2008;119:607–21. taxonomic implications. Biologia. 2012;67:867–74. 191. Salmaki Y, Zarre S, Jamzad Z, Bräuchler C. Trichome micromorphology of 163. Scheen A-C, Lindqvist C, Fossdal CG, Albert VA. Molecular Phylogenetics of Iranian Stachys (Lamiaceae) with emphasis on its systematic implication. tribe Synandreae, a North American lineage of lamioid mints (Lamiaceae). Flora. 2009;204:371–81. Cladistics. 2008;32:299–314. 192. Salmaki Y, Zarre S, Lindqvist C, Heubl G, Bräuchler C. Comparative leaf 164. Roy T, Catlin NS, Garner DMG, Cantino PD, Scheen A-C, Lindqvist C. anatomy of Stachys (Lamiaceae: Lamioideae) in Iran with a discussion on its Evolutionary relationships within the lamioid tribe Synandreae (Lamiaceae) subgeneric classification. Plant Syst Evol. 2011;294:109–25. based on multiple low-copy nuclear loci. PeerJ. 2016;4:e2220. 193. Li HW, Hedge IC. Lamiaceae. In: WU Z-Y, Raven PH, editors. Flora of China, 165. WCSP. World Checklist of Selected Plant Families. Facilitated by the Royal Vol. 17. Beijing: Science Press; St. Louis: Missouri Botanical Garden Press; Botanic Gardens, Kew. 2020. http://wcsp.science.kew.org/. Accessed 16 Nov 1994. pp. 50–299. 2020. 194. Ryding O. Pericarp structure and phylogeny of the Phlomis group 166. Linnaeus C. Species Plantarum, vol. 2. Stockholmiae: Salvius; 1753. (Lamiaceae subfam. Lamioideae). Bot Jahrb Syst. 2008;127:299–316. 167. Dumortier BC. Labiatae. In: Dumortier BC, editor. Florula Belgica. Tournay: J. 195. Seyedi Z, Salmaki Y. Trichome morphology and its significance in the Casterman; 1827. p. 42–9. systematics of Phlomoides (Lamiaceae; Lamioideae; Phlomideae). Flora. 2015; 168. Bentham G. Labiatae. In: De Candolle A, editor. Prodromus systematis 213:40–8. naturalis regni vegetabilis. Vol. 12. Paris: V. Masson; 1848. p. 27–603. 196. Patzak VA. Revision der Gattung Ballota Section Ballota. Ann Naturhist Mus 169. Reichenbach L. Labiatae. In: Reichenbach L, editor. Flora Germanica Wien. 1958;62:57–86. Excursoria. Leipzig: Carolum Cnobloch; 1830–1832. pp. 305–335. 197. Sebald O. Die Gattung Otostegia Bentham (Labiatae) in Afrika und auf der 170. Cosson E, German E. Tribu V. Lamioideae. In: Cosson E, German E, editors. arabischen Halbinsel. Stuttgart Beitr Naturk Ser A. 1973;263:1–84. Synopsis analytique de la Flore des environs de Paris ou description 198. Seybold VS. Revision der Persischen Marrubium-Arten (Labiatae). Stuttgart abregee des familles et des genres. Paris: Fortin, Masson et Cie; 1845. pp. Beitr Naturk, Ser A. 1978;310:1–31. 122–127. 199. Scheen A-C, Albert VA. Molecular Phylogenetics of the Leucas Group 171. Boissier E. Flora Orientalis. Vol. 4(2). Genevae & Basilea: H. Georg; 1879. (Lamioideae; Lamiaceae). Syst Bot. 2009;34:173–81. 172. Bhattacharjee R. Taxonomic studies in Stachys: II. A new infrageneric 200. Kilian N, Hein P, Hubaishan MA. New and noteworthy records for the flora classification of Stachys L. Notes Roy Bot Gard Edinburgh. 1980;38:65–96. of Yemen, chiefly of Hadhramout and Al-Mahra. Willdenowia. 2002;32:239– 173. Tomas-Bárberán FA, Gil MI, Ferreres F, Tomas-Lorente F. Flavonoid p- 69. coumaroylglucosides and 8-hydroxyflavone allosylglucosides in some 201. Ryding O. Reconsideration of Wiedemannia and notes on the labiatae. Phytochemistry. 1992;31:3097–102. circumscription of Lamium (Lamiaceae). Bot Jahrb Syst. 2003;124:325–35. 174. Giuliani C, Bini LM. Glandular trichomes as further differential characters 202. Bendiksby M, Brysting AK, Thorbek L, Gussarova G, Ryding O. Molecular between Stachys subgenus Betonica (L.) Bhattacharjee and Stachys phylogeny and taxonomy of the genus Lamium L. (Lamiaceae): subgenus Stachys. Plant Biosyst. 2012;146:1–8. Disentangling origins of presumed allotetraploids. Taxon. 2011;60:986–1000. 175. Krestovskaya T. Synopsis of the genus Betonica L. (Lamiaceae). Novosti Sist 203. Krawczyk K, Korniak T, Sawicki J. Taxonomic status of Galeobdolon luteum Vyssh Rast. 2014;45:96–109. Huds. (Lamiaceae) from classical taxonomy and phylogenetics perspectives. 176. Lazkov GA, Sennikov AN. Taxonomic corrections and new records in Acta Biol Craco Ser B. 2013;55:18–28. vascular plants of Kyrgyzstan, 4. Mem Soc. Fauna Flora Fenn. 2015;91:67–83. 204. Krawczyk K, Szczecińska M, Sawicki J. Evaluation of 11 single-locus and 177. Ball PW. Stachys L. In: Tutin TG, Heywood VH, Burges NA, Moore DM, seven multilocus DNA barcodes in Lamium L. (Lamiaceae). Mol Ecol Resour. Valentine DH, Walters SM, et al., editors. Flora Europaea. Cambridge: 2013;14:272–85. Cambridge University Press; 1972. p. 151–7. 205. Krestovskaya TV. Menitskia, a new genus of the Labiatae family. Bot Zhurn. 178. Visiani R. Labiatae Juss. In: Visiani R, editor. Flora Dalmatica, Vol. 2. Lipsae: 2006;91:1892–4. Apud Pribdbricum Hofmbistbr; 1847. p. 182–225. 206. Zhao F, Peng H. The complete chloroplast genome of Caryopteris incana 179. Townsend CC. Galeopsis L. In: Tutin TG, Heywood VH, Burges NA, Moore (Lamiaceae) and phylogenetic analysis. Mitochondrial DNA B. 2020;5:1399–400. DM, Valentine DH, Walters SM, et al., editors. Flora Europaea, Vol. 3: 207. Zhao Y, TO T, Xiang C-L. The complete chloroplast genome of Phlomoides Diapensiaceae to Myoporaceae. Cambridge: Cambridge University Press; betonicoides (Lamiaceae), a traditional Tibetan medicinal herb. Mitochondrial 1972. p. 145–7. DNA B. 2020;5:75–6. 180. Meusel H, Jäger EJ, Rauschert SW, Weiner E. Vergleichende Chorologie der 208. Xu W-Q, Losh J, Chen C, Li P, Wang R-H, Zhao Y-P, et al. Comparative zentraleuropäischen Flora. Text u. Karten. Bd. 2. Jena: VEB Fischer; 1978. genomics of figworts (Scrophularia, Scrophulariaceae), with implications for 181. Goldblatt P, Johnson DE. Index to plant chromosome numbers 2001–2003. the evolution of Scrophularia and Lamiales. J Syst Evol. 2019;57:55–65. Monogr Syst Bot Missouri Bot Gard. 2006;106:1–242. 209. Doyle JJ, Doyle JL. A rapid DNA isolation procedure for small amounts of 182. Tomás-Bárberán FA, Gil MI, Ferreres F, Tomás-Lorente F. Correlations fresh leaf tissue. Phytochem Bull. 1987;19:11–5. between flavonoid composition and infrageneric taxonomy of some 210. Andrews S. Fast QC: A quality control tool for high throughput sequence European Galeopsis species. Phytochemistry. 1991;30:3311–4. data. v. 0.10.0. 2010. http://www.bioinformatics.babraham.ac.uk/projects/ 183. Müntzing A. Outlines to a genetic monograph of the genus Galeopsis with fastqc. special reference to the nature and inheritance of partial sterility. Hereditas. 211. Bankevich A, Nurk S, Antipov D, Gurevich AA, Dvorkin M, Kulikov AS, et al. 1930;13:185–341. SPAdes: a new genome assembly algorithm and its applications to single 184. Bendiksby M, Tribsch A, Borgen L, Trávníček P, Brysting A. Allopolyploid cell sequencing. J Comput Biol. 2012;19:455–77. origins of the Galeopsis tetraploids – revisiting Müntzing’s classical textbook 212. Kearse M, Moir R, Wilson A, Stones-Havas S, Cheung M, Sturrock S, et al. example using molecular tools. New Phytol. 2011;191:1150–67. Geneious Basic: an integrated and extendable desktop software platform for 185. Lindqvist C, Motley TJ, Jeffrey JJ, Albert VA. Cladogenesis and reticulation in the organization and analysis of sequence data. Bioinformatics. 2012;28: the Hawaiian endemic mints (Lamiaceae). Cladistics. 2003;19:480–95. 1647–9. 186. Salmaki Y, Heubl G, Weigend M. Towards a new classification of tribe 213. Chen H, Zhang J, Yuan G, Liu C. Complex Interplay among DNA modification, Stachydeae (Lamiaceae): naming clades using molecular evidence. Bot J noncoding RNA expression and protein-coding RNA expression in Salvia Linn Soc. 2019;190:345–59. miltiorrhiza chloroplast genome. Plos One. 2014;9:e99314. 187. Cantino PD, De Queiroz K. International code of phylogenetic 214. Schattner P, Brooks AN, Lowe TM. The tRNAscan-SE, snoscan and snoGPS nomenclature, version 4c. 2010. https://www.ohio.edu/phylocode/. web servers for the detection of tRNAs and snoRNAs. Nucleic Acids Res. Accessed 1 May 2020. 2005;33:686–9. 188. Abu-Asab MS, Cantino PD. Pollen morphology in subfamily Lamioideae (Labiatae) 215. Lohse M, Drechsel O, Kahlau S, Bock R. Organellar Genome DRAW-a suite of and its phylogenetic implications. In: Harley R, Reynolds T, editors. Advances in tools for generating physical maps of plastid and mitochondrial genomes Labiatae science. London: Royal Botanic Gardens, Kew; 1992. p. 361–79. and visualizing expression datasets. Nucleic Acids Res. 2013;41:W575–81. 189. Salmaki Y, Jamazad Z, Zarre S, Bräuchler C. Pollen morphology of Stachys 216. Langmead B, Salzberg SL. Fast gapped-read alignment with Bowtie 2. Nat (Lamiaceae) in Iran and its systematic implication. Flora. 2008;203:627–39. Methods. 2012;9:357–9. Zhao et al. BMC Biology (2021) 19:2 Page 27 of 27

217. Katoh K, Standley DM. MAFFT multiple sequence alignment software version 7.0: improvements in performance and usability. Mol Biol Evol. 2013; 30:772–80. 218. Müller K, Müller J, Quandt D. PhyDe: Phylogenetic data editor, version 0. 9971. 2010. http://www.phyde.de. 219. Talavera G, Castresana J. Improvement of phylogenies after removing divergent and ambiguously aligned blocks from protein sequence alignments. Syst Biol. 2007;56:564–77. 220. Darriba D, Taboada GL, Doallo R, Posada D. JModelTest 2: more models, new heuristics and parallel computing. Nat Methods. 2012;9:772. 221. Ronquist F, Teslenko M, Van der Mark P, Ayres DL, Darling A, Höhna S, et al. MrBayes 3.2: Efficient Bayesian phylogenetic inference and model choice across a large model space. Syst Biol. 2012;61:539–42. 222. Miller MA, Pfeiffer W, Schwartz T. Creating the CIPRES Science Gateway for inference of large phylogenetic trees. Gateway Computing Environments Workshop (GCE). 2010. p. 1–8. 223. Rambaut A, Suchard MA, Drummond AJ. Tracer v1.6. 2014. http://beast.bio. ed.ac.uk/Tracer. 224. Rambaut A. FigTree, v.1.4.2. 2014. http://tree.bio.ed.ac.uk/software/figtree/. Accessed 16 Dec 2019. 225. Stamatakis A. RAxML version 8: A tool for phylogenetic analysis and post- analysis of large phylogenies. Bioinformatics. 2014;30:1312–3. 226. Zhao F, Chen YP, Salmaki Y, Drew BT, Wilson TC, Scheen AC et al. An updated tribal phylogeny of Lamiaceae: evidence from complete plastome sequences. Treebase. http://treebase.org/treebase-web/phylows/study/TB2: S26639?x-access-code=bb02a4c5bc226f4604690ea0f21ccd41&format=html) Accessed 16 July 2020. 227. Thiers B. Index Herbariorum: a global directory of public herbaria and associated staff. New York Botanical Garden’s Virtual Herbarium. 2020. Available at http://sweetgum.nybg.org/ih/. Accessed 16 July 2020.

Publisher’sNote Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.